no
real
Is my toddler a stochastic parrot?
351–360 of 363 posts
Re: Is my toddler a stochastic parrot?
#352Earlier quoted context omitted.
> But life they are not. At best, they are mimics A mimicry of fire is not fire because fire is a reaction between matter. A mimicry of intelligence may well be intelligence, because intelligence is just information processing. We're not talking here about "life", but intelligence and "internal life", aka feelings, thoughts, etc. Edit: as for the argument that "LLMs are simply tools", it has been quite easy for our a…
I havent read it, but Blindsight did a very good job of explaining LLMs. https://en.wikipedia.org/wiki/Blindsight_(Watts_novel) I am saying that LLMs DONT process information, not the way people are implying. The best I can suggest is to please try building with an LLM. Its very hard to explain the nuance, and I am not the only one struggling with it.
You gotta read it!
Re: Is my toddler a stochastic parrot?
#353Earlier quoted context omitted.
> LLMs do not feel or think or experience anything, in the same way that particle physics is not inherently sentient Panpsychists would like a word. > but Markov Chains and LISP bots were no closer to achieving "internal life" or what have you. I'm simply making clear that this is a statement of belief as opposed to a statement of fact . You believe that true life or true sentience has certain qualities that you beli…
> I'm simply saying that we don't know enough about sentience to turn such statements of belief into statements of fact We do! We can look at rocks and very clearly define them as non-sentient. If you disagree with that, you are not deferring to the overwhelming wealth of evidence that can prove rocks are not similar to us. Animists and panpsychics be damned, you only need thermodynamics to prove that we are not the…
> you are not deferring to the overwhelming wealth of evidence that can prove rocks are not similar to us
"Similarity to us" is not a comprehensive metric for sentience, that's the point.
> firmly believe that we can extend simple logic to conclude that LLMs are software
Yes, LLMs are software, so what? Whatever sentience is, it's an integral part of our information processing, and software does information processing. There is literally no reason to conclude that software cannot do the type of information processing that our brain does for sentience.
Even if we agree that panpsychism is incorrect, I still do not agree that we know enough about sentience to claim that software cannot be sentient. I do not claim that LLMs are sentient, but they certainly could be, and the fact that they seem very intelligent, more intelligent than most humans in some ways, is a big red flag.
Re: Is my toddler a stochastic parrot?
#354Earlier quoted context omitted.
> My internal life is related to the structure of my body in an unknown way That's right, an unknown way. So we know the internal structures of computers and their programs, but we don't know the internal structure of humans and how that leads to sentience, but somehow this translates into having good reasons for thinking the relationships in computers are not isomorphic to or meaningfully related to the unknown stru…
> That's right, an unknown way. So we know the internal structures of computers and their programs, but we don't know the internal structure of humans and how that leads to sentience, but somehow this translates into having good reasons for thinking the relationships in computers are not isomorphic to or meaningfully related to the unknown structures that lead to human sentience? Don't you see the big blank there? Th…
> It absolutely follows.
You changed the deduction I described thus moving the goalposts. I'm not going to argue a point I never made. Suffice it to say, the only people assuming anything about LLMs are the ones claiming they're not sentient.
> It would be unbelievably weird if we reverse engineered consciousness without understanding how our bodies generate consciousness
I find it weird that you think this would be weird. Most discoveries are like this. We mastered fire before we understood chemistry, we mastered flight before we understood how birds fly, we have a booming pharmaceutical industry and we still don't really know how most drugs work. Theory and formal understanding often follow discovery and practical applications, not the other way around.
Consciousness and/or sentience will be created before we understand it, and we will only retrospectively recognize that fact. We may have already created it in fact.
Re: Is my toddler a stochastic parrot?
#355Earlier quoted context omitted.
> I'm simply saying that we don't know enough about sentience to turn such statements of belief into statements of fact We do! We can look at rocks and very clearly define them as non-sentient. If you disagree with that, you are not deferring to the overwhelming wealth of evidence that can prove rocks are not similar to us. Animists and panpsychics be damned, you only need thermodynamics to prove that we are not the…
> you are not deferring to the overwhelming wealth of evidence that can prove rocks are not similar to us "Similarity to us" is not a comprehensive metric for sentience, that's the point. > firmly believe that we can extend simple logic to conclude that LLMs are software Yes, LLMs are software, so what? Whatever sentience is, it's an integral part of our information processing, and software does information processin…
Fair enough. I just flat-out disagree; nothing I've seen suggests that software "could be" sentient. The burden of proof is yours, if you want to state that software has the potential for sentience.
Re: Is my toddler a stochastic parrot?
#356Earlier quoted context omitted.
If I'm being honest, I do know they get annoyed by that stuff but I've never really understood why. It's a somewhat common pattern in Mathematics as an avenue for hypotheses to take an existing phenomenon, model some subset of its capabilities, use that to define a new class of behaviour, follow that through to conclusions, then use that to go back to seeing if those conclusions apply to the original phenomenon. A th…
It's an ego thing and a focus thing. They put so much effort into studying the biological details, that they resent people exteacting value without going to that effort, and feel offended at the implication that what they are studying is simple. Ultimately it's largely down to misapprehension of the difference between emulating a neuton and simulating a neuron, and defensiveness about an approximate model.
Hmm, in my experience this is not the primary reason, and I would also argue that the original statement is wrong. Computational neuroscientist are very comfortable (too comfortable perhaps) with simplified models of neurons and brain circuits. We accept them as stick figures. DNNs fit into the typical mindset of synaptic and neural circuits. Reading Synaptic Organization of the Brain should leave you stunned at the relative simplicity of how neuroscientists think the brain works.
Most neuroscientists do not indulge in deep thought because it does not pay the bills. NIH does not push us toward “understanding cognition and consciousness”. They push us toward treatments for Alzheimer’s, schizophrenia, stroke, epilepsy, and addiction. We do not have time for the fun stuff until we are ready to retire, which is when many of us become terrible neurophilosophers, e.g., Eccles.
Re: Is my toddler a stochastic parrot?
#357Earlier quoted context omitted.
> That's right, an unknown way. So we know the internal structures of computers and their programs, but we don't know the internal structure of humans and how that leads to sentience, but somehow this translates into having good reasons for thinking the relationships in computers are not isomorphic to or meaningfully related to the unknown structures that lead to human sentience? Don't you see the big blank there? Th…
> It absolutely follows. You changed the deduction I described thus moving the goalposts. I'm not going to argue a point I never made. Suffice it to say, the only people assuming anything about LLMs are the ones claiming they're not sentient. > It would be unbelievably weird if we reverse engineered consciousness without understanding how our bodies generate consciousness I find it weird that you think this would be…
Great dialog. Thanks to both of you. The boundaries we are trying to define are becoming more and more blurred. In my view the brain is a tool to improve fitness and consciousness/self-consciousness is just a relatively simple higher order recursive system that helps us model the behaviors of others, and spill out long sentences like this one—-nothing too special and nothing that a tweaked LLM with a set of good supervisor algorithms to modulate attention (and perhaps move a body) could not manage.
Human exceptionalism is a slow dying myth.
Re: Is my toddler a stochastic parrot?
#358Earlier quoted context omitted.
> it can mimic human behavior but doesn't have an internal life, feelings, emotions, and isn't conscious I'm curious how you know this. Certainly an LLM doesn't have human internal life, but to claim it has no internal life exceeds our state of knowledge on these topics. We simply lack a mechanistic model of qualia from which we can draw such conclusions.
An LLM is math. It outputs text. Those things aren't alive, and both the software and hardware used to facilitate it is artificial. Once you get that out of the way, sure, I guess it could be "alive" per the same loose definition that any electrical system can exist in an actuated state. It's software, though. I don't think it's profound or overly confident to say that we very clearly know these systems are inanimate…
Why isn’t a brain “math” if you mean a system with which to process input streams to generate adaptive/useful output?
Re: Is my toddler a stochastic parrot?
#359Earlier quoted context omitted.
I havent read it, but Blindsight did a very good job of explaining LLMs. https://en.wikipedia.org/wiki/Blindsight_(Watts_novel) I am saying that LLMs DONT process information, not the way people are implying. The best I can suggest is to please try building with an LLM. Its very hard to explain the nuance, and I am not the only one struggling with it.
I understand the nuance, I think you're just wrong. LLMs augmented with memory are Turing complete [1], so they do process information. This distinction you're trying to draw is a mirage. [1] https://pub.towardsai.net/llms-and-memory-is-definitely-all-...
Augmented with memory and perhaps a bit more—-recursive control of their own attentional mechanisms.
Re: Is my toddler a stochastic parrot?
#360Earlier quoted context omitted.
> you are not deferring to the overwhelming wealth of evidence that can prove rocks are not similar to us "Similarity to us" is not a comprehensive metric for sentience, that's the point. > firmly believe that we can extend simple logic to conclude that LLMs are software Yes, LLMs are software, so what? Whatever sentience is, it's an integral part of our information processing, and software does information processin…
Fair enough. I just flat-out disagree; nothing I've seen suggests that software "could be" sentient. The burden of proof is yours, if you want to state that software has the potential for sentience.
The burden of proof has already been met. The Bekenstein Bound already tells us that any finite volume contains finite information, and any finite system can therefore be perfectly reproduced by a finite state machine. The human brain and body are finite volumes, therefore they can be perfectly replicated by a finite state machine.
Unless you believe in some form of dualism, this reproduction necessarily includes sentience. Therefore, software can be sentient.