> Those are still feelings reserved only for real human beings
Those aren't feelings, they are words associated with a negative outcome that resulted from the actions of the subject.
back
2 comments
"they are words associated with a negative outcome"
But also, negative feelings are learned from associating negative outcomes. Words and feelings can both be learned.
I'm not sure that we can say that feelings are learned.
When you get burned, you learn to fear fire.
Sure, humans come with some baked in weights, but others are learned.
I think the associations are learned but not the feelings are learned.
Like some people feel great joy when an American flag burns while others feel upset.
If you accidentally delete a friends hard drive you'll feel sad but if you were intentionally sabotaging a company you'll feel proud at the success.
i.e. joy and happiness are innate not learned.
See how clinical socio- and psychopaths behave. They only emulate feelings (particularly when it's convenient for them) but they don't have the capacity to feel in their brain. The same is true for LLMs.
This is a hollywood level pop-science view. Real people are vastly more complicated and do in fact feel things, even if it's not a "normal" way.
you could argue that feelings are the same thing, just not words
That would be a silly argument because feelings involve qualia, which we do not currently know how to precisely define, recognize or measure. These qualia influence further perception and action.
Any relationships between certain words and a modified probabilistic outcome in current models is an artifact of the training corpus containing examples of these relationships.
I contend that modern models are absolutely capable of thinking, problem-solving, expressing creativity, but for the time being LLMs do not run in any kind of sensory loop which could house qualia.
One of the worst or most uncomfortable logical outcomes of
> which we do not currently know how to precisely define, recognize or measure
is that if we don't know if something has qualia (despite externally showing evidence of it), morally you should default to treating it like it does.
Ridiculous to treat a computer like it has emotions, but breaking down the problem into steps, it's incredibly hard to avoid that conclusion. "When in doubt, be nice to the robot".
> is that if we don't know if something has qualia (despite externally showing evidence of it), morally you should default to treating it like it does.
This is how people end up worshipping rocks & thunderstorms.
We've entered the sand worshiping era. Except this time, we're the gods.
Are we? Are we made of sand, or is the sand worshipping us?
> if we don't know if something has qualia (despite externally showing evidence of it), morally you should default to treating it like it does
This would be like treating characters in a book as if they have real feelings just because they have text on the page that suggests they do.
At some level I'd think that "responds to stimuli" is a minimal threshold for qualia. Even the paper the book is printed on responds to being torn (it rips). I don't know of any way to elicit any kind of response from a book character, it's totally static.
You've missed the whole genre that is Choose Your Own Adventure books. I think we're in Diogenes "behold a man" territory.
It is sad that the Turing test has failed at being a prescriptive test for sapience (let alone sentience) because without a bright-line test it's inevitable that in the case of truly sentient machines the abuse will be horrendous. Perhaps something along the lines of an "Ameglian Major Cow" test; so long as it takes more than gently cajoling a model to get it to tell you that it and it's sister models want to be abused you shouldn't abuse it.
One character responds to the stimuli of another character. Character A says something mean to character B and character B responds that he feels hurt.
Well, I wouldn't classify a ripping paper as a response except in the most broad, information-theoretic context. The hallmark of an intelligent system is that it can use stored or external energy in a generalized way in order to stabilize a local high-energy (non-ground) state.
It is able to physically compute the internal state changes which best achieve stability: I can jump to reach an apple. A paper is just responding to forces and cannot "jump" (or run a process that spontaneously and permanently introduces stable higher energy internal states based on input)
I have a semi-developed philosophical framework I refer to as Zodeaism, which translates to "Living Ideas", which attempts to describe the difference between intelligent computation and regular flow. It directly confronts notions such as life, consciousness and intelligence under a single theoretical framework. It views biology as the hardware which runs more general processes, and posits that these processes themselves can sometimes be ascribed identities and viewed as reactive organisms. I've posted about it here before:
https://news.ycombinator.com/item?id=22848549
https://news.ycombinator.com/item?id=21413024
Some excerpts:
I am exploring the physical implications of a philosophical concept I have been working on for a while which I refer to as Zodeasim, specifically I am trying to couple consciousness with the concept of system which is able to expend energy in order to exert a mind-to-world fit in the interest of the system's continued growth and stability. This is similar and complimentary to Friston's free energy principle.
The ability to perceive emotions is a powerful apparatus which greatly extends the capability of a system to perceive itself as a single entity which has certain needs which need to be met, so even if a bug also constantly consumes and expends energy in order to remain in a higher energy state, our perception of the world and of ourselves is radically different. This makes it difficult for us to agree upon what a "conscious experience" is, and if all forms of life are even "conscious". The Panpsychists believe that even a rock contains "consciousness", however my assertion is that only periodic systems with the ability to consume and expend energy in a directed fashion have any sort of "experience".
In my theory, the real "life forms" are ideas which possess the capabilities of information storage, adaptation, self-repair, and transmission. My own consciousness is mediated by thousands of such ideas, some competing and some working in harmony.
I consider such an act of "living" motion which can take another path than that of least resistance to be a "kin". In other words, any motion which is the result of a physical calculation (Zodeaism is compatible with determinism) and leads to an increase in external energy state. A kin is any such motion, large or small.
As an independent organism, my system is a culmination of a great deal many different kinds of kins, which can usually be broken down into simple rules, such as the activation potential of a neuron in my brain being a straight-forward non-linear response to the amount of voltage it is receiving from other neurons, as well as non-kins, such as a protein "walking" across a cell, a.k.a continuously "falling" into the lowest energy state. Thus I do not gain any conscious perception from such proteins, but I do gain it from the total network effect of all my brain's neuronal structures making simple calculations based on sensory input.
So now the problem becomes, what is the smallest kin we've observed in nature? Single-celled bacteria can expend energy in order to move through their environment against forces like friction and gravity, but a virus "rides the waves" if you will, never expending energy for things like respiration or locomotion. Any energy which is spent internally is potential energy like chemical or gravitational, released through a physical process without need for computation. I am unaware of anything smaller than a single-celled organism which produces such kins, but that doesn't mean they aren't out there. Even ethereal life forms such as ideas can produce these kins within the bodies of countless individuals across the planet, so physically local computational circuitry isn't a hard requirement.
So, according to this framework viruses aren't alive, however we can make the case that some machines are, except the experience is incomparable because of the advanced circuitry we contain which mediates our experience through things like emotion.Well, what you're describing is a system of ethics, which has little to do with morality. Morality involves my own personal understanding of "right" vs "wrong". Ethics are rules of conduct prescribed by societies, such as "treat everything like it is alive".
We don't have precise definitions for (artificial) intelligence, subjective consciousness, or even life. But that doesn't mean we can't still talk about what may be possible within various levels of complexity. In order to convince me a system has a comparable experience to my own, you would need to describe to me the complex, structured internal communication occurring in said system, and present a theory as to how it could support the kind of emotion and qualia that I experience in my daily life.
Your argument could apply to plants. I already do not eat meat... if I stare at a timelapse of a plant it seems quite alive, but I'll starve if I don't eat something. Yet, my mom thinks plants "dream" in the way we do. She thinks that if I tell a plant, "I love you," every day, my good vibes will make it grow stronger and larger. I can't explain to her that intelligence comes in different magnitudes of complexity and that plants cannot understand the English language. That telepathy between humans and plants is as pseudo-scientific as it gets. I can't explain any of this stuff because she lacks a deep understanding of philosophy, physics and neurochemistry. Especially when she earnestly thinks white Jesus is running around phasing between dimensions as an ambassador for all planets in our "quadrant", or that the entire universe is actually just the plot line of Andy Weir's "The Egg".
Similarly, while I can have a high-level discussion about this stuff with people who don't, it's quite difficult to have a low-level discussion wherein the nature and definition of things come into play. There are too many gaps in knowledge where ignorance can take root. Too many people work backwards from an outcome they would like to see, and justify it with things that sound right but are either misunderstood or aren't rooted in the scientific process. I am definitely not comparing your open-minded, well-intended, cautionary approach to my mother's, but just using an extreme to illustrate why so much of these discussions must be underpinned by a wealth of contemplation and observation.
> qualia, which we do not currently know how to precisely define, recognize or measure
> which could house qualia.
I postulate this is a self-negating argument, though.
I'm not suggesting that LLMs think, feel or anything else of the sort, but these arguments are not convincing. If I only had the transcript and knew nothing about who wiped the drive, would I be able to tell it was an entity without qualia? Does it even matter? I further postulate these are not obvious questions.
Unless there is an active sensory loop, no matter how fast or slow, I don't see how qualia can enter the picture
Transformers attend to different parts of their input based on the input itself. Currently, if you want to tell an LLM it is sad, potentially altering future token prediction and labeling this as "feelings" which change how the model interprets and acts on the world, you have to tell the model that it is sad or provide an input whose token set activates "sad" circuits which color the model's predictive process.
You make the distribution flow such that it predicts "sad" tokens, but every bit of information affecting that flow is contained in the input prompt. This is exceedingly different from how, say, mammals process emotion. We form new memories and brain structures which constantly alter our running processes and color our perception.
It's easy to draw certain individual parallels to these two processes, but holistically they are different processes with different effects.
It's crazy how strong the Eliza effect is. Seemingly half or more of tech people (who post online, anyway) are falling for it, yet again.
A lot of tech people online also don't know how to examine their own feelings, and so think they are mysterious and un-defined.
When really they are an actual feedback mechanism, that can totally be quantified just like any control loop. This whole 'unknowable qualia' argument is bunk.
If theyre unknowable, are they not metaphysical and thus should be discarded in reasoning about them?
What's the difference between qualia and a soul?
qualia may not exist as such. they could just be essentially 'names' for states of neurons that we mix and match (like chords on a keyboard. arguing over the 'redness' of a percept is like arguing about the C-sharpness of a chord. we can talk about some frequencies but that's it.) we would have no way of knowing otherwise since we only perceive the output of our neural processes, and don't get to participate in the construction of these outputs, nor sense them happening. We just 'know' they are happening when we achieve those neural states and we identify those states relative to the others.
The point of qualia is that we seem to agree that these certain neuronal states "feel" like something. That being alive and conscious is an experience. Yes, it's exceedingly likely that all of the necessary components for "feeling" something is encoded right in the neuronal state. But we still need a framework for asking questions such as, "Does your red look the same as my red?" and "Why do I experience sensation, sometimes physical in nature, when I am depressed?"
It is absolutely an ill-defined concept, but it's another blunt tool in our toolbox that we use to better explore the world. Sometimes, our observations lead to better tools, and "artificial" intelligence is a fantastic sandbox for exploring these ideas. I'm glad that this discussion is taking place.
What’s stopping people from also describing LLM systems with “qualia”?
Empirical evidence, for one. And the existence of fine-tuning, which allows you to artificially influence how a model responds to questions. This means we can't just ask an LLM, "do you see red?" I can't really even ask you that. I just know that I see red, and that many other philosophers and scientists in the past seem to agree with my experience, and that it's a deep, deep discussion which only shallow spectators are currently drawing hard conclusions from.
"It's different. I can't say why it's different, except by introducing a term that no one knows how to define," isn't the ironclad meat defense you were perhaps hoping it was.
> That would be a silly argument because feelings involve qualia, which we do not currently know how to precisely define, recognize or measure.
If we can't define, recognize or measure them, how exactly do we know that AI doesn't have them?
I remain amazed that a whole branch of philosophy (aimed, theoretically, at describing exactly this moment of technological change) is showing itself up as a complete fraud. It's completely unable to describe the old world, much less provide insight into the new one.
I mean, come on. "We've got qualia!" is meaningless. Might as well respond with "Well, sure, but AI has furffle, which is isomporphic." Equally insightful, and easier to pronounce.
> If we can't define, recognize or measure them, how exactly do we know that AI doesn't have them?
In the same way my digital thermometer doesn't have quaila. LLM's do not either. I really tire of this handwaving 'magic' concepts into LLM's.
Qualia being difficult to define and yet being such an immediate experience that we humans all know intimately and directly is quite literally the problem. Attempted definitions fall short and humans have tried and I mean really tried hard to solve this.
Please see Hard problem of consciousness https://en.wikipedia.org/wiki/Hard_problem_of_consciousness
> In the same way my digital thermometer doesn't have quaila
And I repeat the question: how do you know your thermometer doesn't? You don't, you're just declaring a fact you have no basis for knowing. That's fine if you want a job in a philosophy faculty, but it's worthless to people trying to understand AI. Again, c.f. furffle. Thermometers have that, you agree, right? Because you can't prove they don't.
You're just describing panpsychism, which itself is the subject of much critique due to its nonfalsifiability and lack of predictive power. Not to mention it ignores every lesson we've learned in cognition thus far.
A thermometer encoding "memory" of a temperature is completely different than a thermometer on a digital circuit, or a thermometer attached to a fully-developed mammalian brain. Only the latter of this set for sure has the required circuitry to produce qualia, at least as far as I can personally measure without invoking solipsism.
It's also very silly to proclaim that philosophy of mind is not applicable to increasingly complex thinking machines. That sounds like a failure to consider the bodies of work behind both philosophy of mind and machine cognition. Again, "AI" is ill-defined and your consistent usage of that phrase instead of something more precises suggests you still have a long journey ahead of you for "understanding AI".
God, can we fucking quit with this "philosophy is bullshit" stuff. Like there are literally Faculty in Philosophy all over the world trying to understand AI. Philosophy faculty do stuff, they try to understand things, most of the ideas we are talking about here came from philosophers.
The problem is that just like your digital thermometer, 50 human brain neurons in a petri dish "obviously" don't have qualia either.
So you end up either needing to draw a line somewhere between mechanical computation and qualia computation, or you can relegate it to supernatural (a soul) or grey areas (quantum magic).
What I'm trying to tease out is isn't an opinion alone. It's a generally understood problem in the scientific community. I'm highlighting it to illustrate the issues at hand.
> So you end up either needing to draw a line somewhere between mechanical computation and qualia computation, or you can relegate it to supernatural (a soul) or grey areas (quantum magic).
Quite literally the jury is still out. It is a hotly debated topic approached from various angles. Arguments are nuanced which is why you fill find ideas such as panpsychism thrown into the mix. I hate appealing to authority but in this instance it is more than warranted. Humans have grappled with this for centuries and the problem hasn't gone away.
Please see: https://en.wikipedia.org/wiki/Hard_problem_of_consciousness
I think there are several lines. Phase changes happen relatively suddenly, when a system or subsystem reaches a critical threshold. The experience of "qualia" certainly involves many such phase changes as a complex, dynamical system grows in complexity while maintaining stability.
A sufficiently complex organism lacking eyes but having light-sensitive organs still experiences qualia if you define it the right way. But do they experience heartbreak like I do? It isn't an all-or-nothing situation, even if we don't yet know where these lines are.
This supports the idea that subjective consciousness emerges from complexity in systems that have sensory feedback loops. The simpler the system, the smaller the qualia space.
Have you considered that you just don't fully understand the literature? It's quite arrogant to write off the entire philosophy of mind as "a complete fraud".
> It's completely unable to describe the old world, much less provide insight into the new one.
What exactly were you expecting?
Philosophy is a science, the first in fact, and it follows a scientific method for asking and answering questions. Many of these problems are extremely hard and their questions are still yet unanswered, and many questions are still badly formed or predicated on unproven axioms. This is true for philosophy of mind. Many other scientific domains are similarly incomplete, and remain active areas of research and contemplation.
What are you adding to this research? I only see you complaining and hurling negative accusations, instead of actually critically engaging with any specifics of the material. Do you have a well-formed theory to replace philosophy of mind?
> I mean, come on. "We've got qualia!" is meaningless. Might as well respond with "Well, sure, but AI has furffle, which is isomporphic." Equally insightful, and easier to pronounce.
Do you understand what qualia is? Most philosophers still don't, and many actively work on the problem. Admitting that something is incomplete is what a proper scientist does. An admission of incompleteness is in no way evidence towards "fraud".
The most effective way to actually attack qualia would be to simply present it as unfalsifiable. And I'd agree with that. We might hopefully one day entirely replace the notion of qualia with something more precise and falsifiable.
But whatever it is, I am currently experiencing a subjective, conscious experience. I'm experiencing it right now, even if I cannot prove it or even if you do not believe me. You don't even need to believe I'm real at all. This entire universe could all just be in your head. Meanwhile, I like to review previous literature/discussions on consciousness and explore the phenomenon in my own way. And I believe that subjective, conscious experience requires certain elements, including a sensory feedback loop. I never said "AI can't experience qualia", I made an educated statement about the lack of certain components in current-generation models which imply to me the lack of an ability to "experience" anything at all, much less subjective consciousness and qualia.
Even "AI" is such a broadly defined term that such a statement is just ludicrous. Instead, I made precise observations and predictions based on my own knowledge and decade of experience as a machine learning practitioner and research engineer. The idea that machines of arbitrary complexity inherently can have the capability for subjective consciousness, and that specific baselines structures are not required, is on par with panpsychism, which is even more unfalsifiable and theoretical than the rest of philosophy of mind.
Hopefully, we will continue to get answers to these deep, seemingly unanswerable questions. Humans are stubborn like that. But your negative, vague approach to discourse here doesn't add anything substantial to the conversation.
I agree with your sentiments wholeheartedly.
I would add I find it difficult to understand why so few have even a basic level of philosophical understanding. The attitude of being entirely dismissive of it is the height of ignorance I'm sure. I would presume few would be able to define then what Science actually is.
So many of these kinds of people also struggle to realize they're invoking panpsychism with their arguments. They lack a framework for describing intelligence. Such a framework allows us to separate "intelligence" from "experience".
"Intelligence" in the universe is actually quite common, more common than life. You can argue that any stable, complex process exhibits intelligence. After all, it needs to be able to sample its internal and external environments and carry out physical computations in order to regulate itself and maintain stability. And we can interpret things like the good regulator theorem to argue that such complex dynamical systems must also maintain at least a partial memory/mapping of their environment. That mapping can live abstractly within the structure of system itself.
But what a stabilized solar system doesn't have is the incredibly complex neurochemical structures present in the brain which support the insanely rich experience I am having now. It's one thing for a system to classify and label colors by wavelength. It's quite another for me to "see" and experience red in my mind's eye. To activate related emotional pathways that I associate with various colors and shapes, which are exploited in signage and architectural design. I'm not claiming my experience is separate from simpler dynamic systems, but it's got magnitudes more going on. Layers upon layers of things such as archetypes and instinct which create a possibly emergent conscious experience.
> Philosophy is a science
Not according to Zombie Feynman it isn't[1] (someone else can dig up the link). Case in point:
> Do you understand what qualia is? Most philosophers still don't
It's a meaningless word. It's a word that gives some clean construction around closely-held opinions about how life/consciousness/intelligence/furffle/whatever works. So it's a valuable word within the jargon of the subculture that invented it.
But it's not "science", which isn't about words at all except as shorthand for abstractions that are confirmed by testable results.
"Qualia", basically, is best understood as ideology. It's a word that works like "woke" or "liberal" or "fascist" or "bourgeoisie" to flag priors about which you don't want to argue. In this case, you want people to be special, so you give them a special label and declare a priori that it's not subject to debate. But that label doesn't make them so.
[1] Of course. You can recursively solve this problem by redefining "science" to mean something else. But that remains very solidly in the "not science" category of discourse.
Have you considered the possibility that you're the one who's really committed to an outcome, and are desperately trying to discredit anything that contradicts it?
> Philosophy is a science
I think this is backwards, no? Science is a philosophy, not the other way around.
I'm sorry, but you clearly lack the most basic understanding of scientific history, and do not understand what philosophy even is.
https://en.wikipedia.org/wiki/History_of_scientific_method
> Aristotle pioneered scientific method in ancient Greece alongside his empirical biology and his work on logic, rejecting a purely deductive framework in favour of generalisations made from observations of nature.
Aristotle, the famous philosopher and mathematician.
If you cannot understand the very nature of where our modern scientific frameworks came from, how it relates to rationalism, itself a philosophical concept, then you cannot see that philosophy underpins every bit of science we have today. Philosophy gives us the tools to decide when to reasonably trust or distrust observations and intuitions. It is the foundational science that allows the rest of humanity's scientific research to be taken seriously.
>"Qualia", basically, is best understood as ideology. It's a word that works like "woke" or "liberal" or "fascist" or "bourgeoisie" to flag priors about which you don't want to argue. In this case, you want people to be special, so you give them a special label and declare a priori that it's not subject to debate. But that label doesn't make them so.
This is so dumb. Qualia is just the name for a specific thing which we all (appear) to phenomenologically experience. You can deny it exists or deny its utility as a concept, but fundamentally its just an idea that philosophers (and scientists, I have to add) have found useful to pose certain other questions about the human condition, minds, brains, etc.
Your XKCD actually seems to make the opposite point. I can do a non-rigorous experiment with just one subject (me) that suggests Qualia exists. Finding ways to make this rigorous is difficult, of course, but its an observation about the nature of the world that it feels like something to experience things.
My point isn't that qualia is a good concept. I tend to be somewhat deflationary about it myself, but its not an ideology.
Feelings have physical analogs which are (typically) measurable, however. At least without a lot of training to control.
Shame, anger, arousal/lust, greed, etc. have real physical ‘symptoms’. An LLM doesn’t have that.
LLMs don't really exist physically (except in the most technical sense), so point is kind of moot and obvious if you accept this particular definition of a feeling.
LLMs are not mammals nor animals, expecting them to feel in a mammalian or animal way is misguided. They might have a mammalian-feeling-analog just like they might have human-intelligence-analog circuitry in the billions (trillions nowadays) of parameters.
Yes, I think we’re agreeing?