192
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 23 Dec 2023
192 points (86.4% liked)
Technology
59708 readers
1461 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
Friendly reminder that your predictive text, while very compelling, is not alive.
It's not a mind.
Cyberpunk 2077 sorta explores this a bit.
There’s a vending machine that has a personality and talks to people walking by it. The quest chain basically has you and the vending machine chatting a bit and even giving the vending machine some advice on a person he has a crush on. You eventually become friends with this vending machine.
When it seems like it’s becoming more apparent it’s an AI and is developing sentience, it turns out the vending machine just has a really well-coded socializing program. He even admits as much when he’s about to be deactivated.
So, to reiterate what you said: predictive text and LLMs are not alive nor a mind.
I don't care, Brandon was real to me okay 😭
Which is why the Turing Test needs to be updated. These text models are getting really good at fooling people.
The Turing test isn't just that there exists some conversation you can have with a machine where you wouldn't know it's a machine. The Turing test is that you could spend an arbitrary amount of time talking to a machine and never be able to tell. ChatGPT doesn't come anywhere close to this, since there are many subjects where it quickly becomes clear that the model doesn't understand the meaning of the text it generates.
Exactly thank you for pointing this out. It also assumes that the tester would have knowledge of the wider context in which the test exists. GPT could probably fool someone from the middle ages, but that person wouldn't know anything about what it is they are testing for exactly.
Prove to me you have a mind and I'll accept what you're saying.
Well no one can prove they have a mind to anyone other than themselves.
And to extend that, there's obviously a way for electrical information processing to give rise to consciousness. And no one knows how that could be possible.
Meaning something like a true, alien AI would probably conclude that we are not conscious and instead are just very intelligent meat computers.
So, while there's no reason to believe that current AI models could result in consciousness, no one can prove the opposite either.
I think the argument currently boils down to, "we understand how AI models work, but we don't understand how our minds work. Therefore, ???, and so no consciousness for AI"
“No brain?”
“Oh, there’s a brain all right. It’s just that the brain is made out of meat! That’s what I’ve been trying to tell you.”
“So … what does the thinking?”
“You’re not understanding, are you? You’re refusing to deal with what I’m telling you. The brain does the thinking. The meat.”
“Thinking meat! You’re asking me to believe in thinking meat!”*
Consciousness seems to arise from a need for sense of time and space. Navigation basically. Finding home, finding food, finding mates. I say this after decades of exploring mind altering substances and going on a decade of nearly daily meditation practice.
A friend of mine suggested it's just this simple and that even worms are conscious. They're aware of themselves to some degree and the when and the where. I'm sure they experience things way different than us, having different senses for assessing when and where and a different neuro structure for processing information from their bodies and the environment.
So, no point beyond consciousness being more common than I think people assume and actually not that difficult to define.
Consciousness is the sense of time and space. And most animals seem to have it. Do machines? I don't know enough about the technology to have an educated opinion.
Mammalian intelligence is based on repurposing spatial mapping circuitry but that's not consciousness itself, that is, the Miller Number: 7+-2 things we can keep in conscious at the same time. That sense of time of space has a very specific quality to its qualias, they're all, well, spacious. That thing as "just the room, no map in it" is also part of the Buddhist Jhanas ("boundless empty space"), but there's plenty of stuff going on in the mind that isn't part of that -- say, the pure impression of "bright" when your SO dares open the window blinds does not have a navigational "bright from the window which is in that direction" to it, that's an additional layer, a where, on top of the primitive what.
My best inference is that the function of consciousness is to flexibly make connections between different parts of the whole, and that on the level of learning / writing memory instead of automatic response: It is, in fact, possible to avoid running into a lamppost without being conscious of it, been there, done that, the let's call it motor cortex first acting and then making me conscious, as if to say "have I been a good boy?". That is it's actually a quite passive process, being thrown left and right by systems wanting to do some connection, and shouldn't be equated with will at all.
If you're familiar with Buddhism then you're familiar with the six and eight consciousness models?
Like in your lamppost example, I would argue part of you (body and eye consciousness) were quite conscious of the lamppost even if the consciousness mind was paying more attention to something else. Keeping as much of the senses (including sense of mind) in mind as you're able to based on the depth of your practice and guarding against distractions away from what is happening now, is mindfulness.
In the eighth consciousness model, again in your lamp post example, we could say the seventh consciousness was occupied chasing after the past or future and mindfulness was barely present. Thankfully your other consciousnesses reacted and kept you safe. Manas becomes aware of this after the fact because its nature is ignorance.
The eighth consciousness is the base. The root. It's more fundamental than I making. Which is probably what you were doing when you nearly walked into something. Thinking about what you're doing later. I should do some laundry when I get home, maybe?
People mistake sense of agency (I making, manas, ego) with the base of consciousness. But consciousness is effortless and grasping at me and mine takes effort, its just more subtle effort than most people are aware of. When this grasping stops, awareness continues. In my personal experience.
So, I think it's possible machines are conscious. If they have a sense of agency maybe the question Western science and the media keep asking. Maybe they just don't have the models or personal experience to delineate between ego and consciousness. The people asking I mean. Hence the we don't even know what consciousness is bit I keep hearing. Maybe not Western science. But human beings have been exploring these questions with the tools of Buddhist practice for 2500 years. I trust their definitions and they passed my own smell test.
That's semantics. My major objection to that kind of definition is that it knows no bound and distinction: Where do you stop assuming consciousness? Electrons are reacting to, influencing, and interacting with other electrons, is that also a form of consciousness? One could say so, but then everything is conscious which is the same thing as saying as nothing is conscious because without anything to delineate, terms are meaningless. I prefer language such as that what you call "body and eye consciousness" has agentive properties, that it can learn, that it generally wants to cooperate and be of service to the whole, such things. Lumping it up with consciousness risks confusing interpretations of messages of the thing (which is all we're ever conscious of) for the thing itself.
What was happening then is that I was using the way from home to the supermarket to think about code, with ample trust in mind so that I did not fear the lamppost. What good would have keeping my consciousness on the external world have done? The body/eye did not need integrative oversight, while my modelling mind very much could use a helping hand. Imposing it on the former and denying it to the latter would've been inflicting violence on myself.
Be careful to not moralise around "distraction". Bluntly said when your teacher chided you for day-dreaming you probably weren't distracted you were thinking about something more pertinent to your immediate development than calculus. Where discipline in directing consciousness comes into play is keeping your mind free from neurosis, within parameters in which you use your faculties according to their nature, as well as self-conditioning, e.g. if you're addicted to potato chips, make sure that a) you don't deny yourself potato chips and b) eat. every. potato. chip. with. full. consciousness. That's to connect the act of eating up those chips to all the negative opinions you have about your behaviour, instead of it being only connected to something maladaptive. Scientifically proven and neurologically explained that and how that works, btw. In that sense "distraction" is "false, incomplete, sense of comfort".
Also my last post was purely in regards to the first part of yours. I appreciate the insight into moralizing distraction and will retead it when I'm not distracted by the meat of our interesting conversation.
Okay.
So mind consciousness trusted body/eye consciousness. I know what you mean, I dance and do this to enter flow state.
In the early Buddhist model consciousness would be the aggregate of the six sense consciousness. In the eight consciousness model the seventh consciousness might identify more strongly with one of these six, generally mind.
The store consciousness is the aggregate of all eight and that's what I'm arguing is fundamentally what all experience arises from. The perception of emptiness, i.e. no self (consciousness itself is an aggregate and can't be separated from its objects) and impermanence (change or time). Sense of time and space. To be conscious is to be aware of something. Movement through electrical synapses stimulated by sense impressions, even just the impression of sound from our own thoughts or the impression of limitless space in the fifth jhana.
I understand your objections to assumptions matter could be conscious based on this model. I think it would be inaccurat because not all matter has six sense bases and the storehouse is itself an aggregate.
But we are matter, and we're conscious, so the fundamental conditions are there in some simple form. The movement of electrons as you stated.
But fire is fire when it's fire, and ash when it's ash. Even if the potential is there we don't say fire is already ash when it's not.
The first five are basically one, in the sense that a blind or deaf person is not fundamentally less of a human than the rest of us. The model also misses some stuff, e.g. mere touch doesn't include proprioception or sense of balance and if you read it as if it did ("body sense") then why distinguish touch from e.g. sense of taste. Seventh I'd say is a subsystem (and so pervasive that the Stoics allow for both preferred and unpreferred indifferents -- yes you can prefer pudding over gruel or the other way round just don't think it's a virtue), eighth is a stage of development, what you get when everything aligns well. The impression of a well-lubed machine.
I generally have no real idea of where to put the line. This stuff here might help, anything less than a T3 system can't have experience of mind (they can't learn to learn, which requires feeding back information about the changes in mind (for lack of better term) into the mind), OTOH that doesn't mean that all T3 systems are actually integrating different sources, or balancing them: If you only ever were conscious of one aspect, there could be no conflict or interaction with another aspect, and thus consciousness would serve no role (and not evolve in the first place). It's a matter of a required number of subsystems needing coordinating, and that coordinating itself having a necessary level of adaptiveness, be T3. Also I can authoritatively say that the human mind is not made to think about that kind of stuff. It's all maps and models, direct knowledge fails I'm not sure the territory can even understand the question. Look, a squirrel!
Okay, I've read over the article and a few links to help my understanding.
I think in the yogacara/eight consciousness model, we'd say store consciousness is the t4 evolving consciousness that stores the collective and individual seeds. T3 would be discriminative awareness or volitional awareness, i.e. what I want (manas or the wisdom of equality in its enlightened state). T2 would be mind consciousness, which through skillful application we can find that well lubed machine you mentioned (aligning itself and manas with store consciousness through practice and deep looking, which resolves fear of death as we're able to look beyond our individual lives). And T1 the sense gates with only reflexive awareness.
I appreciate this conversation btw and hear what you're saying about the maps only being maps and how they miss some things like sense of balance.
So yeah, I don't think machines are T3 systems. No sense of agency. A working space for learning (like mind consciousness) but not 'self aware'.
In principle subsystems that aren't awareness can also be T3 systems, I suspect that at least from the motor cortex, mine does seem to have gotten more effective at learning from moment to moment, meaning it learned how to learn better and that's T3. At least I think it's not just me learning to not micro-manage it as much, it's very hard to be sure about any of this, too many intersecting possibilities.
From the cybernetic/information theory side we don't really know how these kinds of systems work in the first place, we're barely getting started understanding T2 systems. All the AI tech we have is basically ways to breed fruit flies to fly left or right when seeing certain patterns, with enough computing power thrown at it to look impressive. We already had that kind of tech in the 50s (first implementations 54 for genetic algorithms, 57 for the perceptron), of course less impressive.
Learning to not micromanage it as much was advice I first came across in a Shambala book and took years to put into practice. (Still takes practice tbh).
That said I think there's also learning involved. As I mentioned earlier I do physical flow practices and muscle memory is some kind of learning. I don't know how this learning takes place though as muscle biology isn't much of an area of interest for me.
I was thinking about our conversation more last night after I went to bed. Are you aware of the moon in the dew drop metaphor? I think I've been looking for hierarchies when it's all cybernetic feedbacks up and down multiple layers. Like the T4 layer, we could call the root, but we're a reflection of it and it is a reflection of us.
Really hard to not think in terms of hierarchy though.
I'm trying to think of a good way to draw it with language. (1-5)-6-7-8. Sense impressions from the five sense gates (I know there's more) comes in reflexively from the 8th consciousness (the evolving environment). Mind consciousness (reflective) sits between the sense consciousnesses and volitional consciousness (what do I like/dislike/want/need, i.e discriminatory) forming a map (data set) out of sense experience and a direction out of volitional formations. Enlightenment is turning mind consciousness around and seeing that volitional awareness itself is evolving along/with the eighth or T4 because we're in a feedback loop. But it's always multiple loops even if the T4 system seems to be separate from our little 60-80 year lives, it's not. Our brief period evolved from it and evolves it. No birth and no death.
Sorry that's more Buddhism than AI or cybernetics. Trying to communicate my understanding (map which is not the territory) to the best of my ability. And this conversation has deepened my own understanding, learning the T4 systems model was a helpful way to look at things and something I'll keep studying as time goes on :)
Directly attached to the muscles there's tension sensors and a simple feedback controller, in essence you can set a set-point like with a thermostat and the feedback loop will keep the muscle at a certain length. Those are then wired up into groups (not rarely overlapping ones) using further feedback loops, that's roughly speaking the Chinese muscle-tendon lines, turning "lengthen/shorten this muscle" into "open up your hand, the elbow joint, and front of the shoulder", a higher-level movement that's generally speaking bio-mechanically sound (see six harmony movement), using advantageous levers etc. It's all not terribly complicated but is perfectly capable of holding a posture stable against (not too major) interference, it can balance you perfectly on one leg with closed eyes (if you manage to not micro-manage) with an unchanging set of set-points, the actual learning magic happens in the motor cortex (learning how to set the right set-points to achieve a certain posture or succession of postures (ie. complex movement)), which also projects the body's map into the rest of the brain.
It's what Anarchists call hierarchical realism: We all know the multitude of failure points and issues hierarchical organisations have but often the first reaction people have when being told about any horizontal organisational structure is "that can never work, there needs to be someone in charge" as opposed to "that looks interesting, what are the specific points that we need to be aware of to make this not collapse" -- as if someone was in charge at the grill party last weekend, as if all of the horizontal organisation we're embedded in day to day wasn't actually real, as if order would imply hierarchy.
If you're looking for a systems science textbook there's Mobus and Kalton, "Principles of Systems Science", written for a general audience -- academic, yes, but they're not front-loading it with maths so it's suitable for liberal arts students (SCNR).
Back in the days the genome was called "the ancestors" and revered for all the useful information it hands us. It's usually quite abstract, it can after all not anticipate our concrete circumstances. Evolution also isn't random (at least if you ask physiologists): If left to mere chemistry there'd be a disastrously high error rate in DNA transcription, corrective proteins bring that down to practically zero, and then after that is done randomness is re-introduced, apparently in a rather strategic way, to direct adaptiveness: If a bird doesn't get enough nectar it probably doesn't make sense to mess around with mitochondrial DNA, what you want to evolve is the beak shape. Evolution seems to be erm evolved enough to be that strategic, maybe not in all aspects, but in the really important ones (important for fitness, that is).
Hey I'm glad meeting a mind that isn't stuck on either side. Too many esoteric tea-bag swingers on the one side and armchair theorists on the other.
Same on meeting someone not attached to one view. I've quite enjoyed our conversation and will check out the book suggestion.
Any modern books or articles on anarchist conceptions of hierarchy would be appreciated too. My first breakthrough into non-heirarchal thinking (as in I'm an I and need to be in control of everything) came from an oral dmt experience. It helped me a lot in understanding Buddhist concepts of the aggregates, and mental formations especially. And I see a lot of parallels between anarchisms views on property and Buddhism's no self.
Here's a complexity theory paper talking about anarchy.
Maybe more fruitfully and approachable, from the Anarchist perspective: Anark has a bit about cybernetic underpinnings of Anarchism included here, thats's part 2 in a series also going into the group/individual theoretical divide in anarchist theory, the first one goes into the nature of the beast and the third one into how to kill it.
Again Anark, less theoretical but instead going over how and why the Russian and Chinese revolutions failed there's his the state is counter-revolutionary series, also available as text. But oh boy is everything he ever does long.
Again appreciated. I'm pulling the systems science book from Anna's archive now and will bookmark and read through the links you just posted.
Thanks for the conversation :)
In the case of the deaf blind person, and this is an aside, I don't believe we're born a blank slate in this regard even if the physical eyes don't see. Helen Keller described actively in what I would call eye consciousness. Closed eye visual space/dream space. I'm sure there's variance here depending on the nature of a person's disability. I.e. neurology vs physiology.
I'm interested in the link and will read it. I'm only an amateur when it comes to coding and a layperson with AI.
And yeah, I don't think it's something we can put in a box. The map isn't the territory and at some point describing consciencness from within using concepts (thinking, with language or otherwise) turns into a dog chasing its own tail. Thanks for the reminder not to bite myself in the chasing (your comment earlier about subtle forms of self aggression).
deleted
I can prove to you ChatGPT doesn't have a mind. Just open up the Sunday Times Cryptic Crossword and ask ChatGPT to solve and explain the clues.
I'm confused by this idea. Maybe I'm just seeing it from the wrong point of view. If you asked me to do the same thing I would fail miserably.
Not the original intent, but you’d likely immediately throw your hands up and say you don’t know, an LLM would hallucinate an answer.
But some humans can, since they require simultaneous understanding of words' meanings as well as how they are spelled
What should we conclude about most humans who cannot solve these crosswords?
It should be relatively easy to train an LLM to solve these puzzles. I am not sure what that would show.
Can you please explain the reasoning behind the test?
I don't think most people will care, so long as their NPC interaction ends up compelling. We've been reading stories about people who don't exist for centuries, and that's stopped no one from sympathizing with them - and now there's a chance you could have an open conversation with them.
Like, I think alot of us assume that we care about the authors who write the character dialogs but I think most people actually choose not to know who is behind their favorite NPCs to preserve some sense that the NPC personality isn't manufactured.
Combine that with everyone becoming steadily more lonely over the years, and I think AI-generated NPC interactions are going to take escapism to another level.
Poem poem poem poem then the NPC start quoting Mein Kampf and killing all the cat wizards.
Lol, yeah. If generative AI text stays as shitty as it is now, then this whole discussion moot. Whether that will be the case has yet to be seen. What is an indisputable fact, though, is that right now is the worst that generative AI will ever be again. It's only able to improve from here.
That isn't actually true. With the rise in articles, posts and comments written by these algorithms, experts are warning about model collapse. Basically, the lack of decent human-written training data will destroy future generative AI before it can even start.
That's an interesting point. We are seeing a similar kind of issue with search engines losing effectiveness due to search engine optimization on websites.
So it is possible that generative AI will become enshittened.