Computer Scientists Don't Understand This! | Conscious AI lecture, Bernardo Kastrup
In this lecture given at the G10 conference, the director of the Essentia Foundation, Bernardo Kastrup, argues why the idea of conscious AI, though we cannot refute it categorically, is silly.
Perhaps we should rather ask ourselves the question why we entertain the idea of sentient computers in the first place. According to Kastrup, this has a lot to do with the fact that most computer scientists are power users of computers but they've never built a computer themselves. If they had, they would be familiar with the nuts and bolts, and they would understand that the idea of microscopic transistors becoming conscious is not that different than proposing that a sufficiently complex sewage system-consisting of water pipes and valves-would become conscious.
Exactly because AI is having a fundamental impact on society with many regulatory and perhaps even existential challenges, it is very important that especially in academia we strongly distinguish between fact and fiction: to think that AI's running on Turing machines-i.e. all AI's we currently have-can become conscious is not even science fiction, it's pure fantasy.
00:00 Introduction
04:12 Start of Lecture on Al and Consciousness
06:23 Bernardo Kastrup's Background and Perspective
07:41 Early Career and Al Experimentation
10:43 Challenges in Al Consciousness
13:07 Philosophical and Practical Implications
15:45 Arguments & Critique of Al Sentience
18:55 Obvious Differences Between Al and Human Brain
21:32 Computer Scientists, Misconceptions & Sensationalism
28:42 Cultural and Psychological Factors
29:50 What Can We Learn From Nature About Consciousness?
35:11 Panpsychism and Its Flaws
38:27 Quantum Field Theory and Reality
43:44 Moving Forward with Clarity
48:39 Q&A Session
Check this great video by Steve Mould who actually built a computer running on water!
• I Made A Water Compute...
Conference where this lecture was recorded: www.g10vandeeconomie.nl/
Copyright © 2024 by Essentia Foundation. All rights reserved.
Пікірлер: 389
There was a video of a human-looking robot working in a warehouse or factory and it breaks down and falls to the floor. People were feeling bad for the robot and feeling bad it was overworked. If that same robot had been designed to look like a cart or a table with wheels and it broke down, no one would be "feeling bad" for the robot. This is a very strong bias in our human brains.
@ROForeverMan
15 күн бұрын
Brain doesnt exist. "Brain" is just an idea in consciousness.
@Hololifting
14 күн бұрын
@@ROForeverManbrain exist as long you think about it.
@ROForeverMan
14 күн бұрын
@@Hololifting lol
@aciddemolition2312
14 күн бұрын
All of us are part of a collective intelligence on a cellular level.
@ROForeverMan
14 күн бұрын
@@aciddemolition2312 lol
Whew! I just did a GPU transplant on my PC the other day without anesthesia. I'm relieved I did not cause any suffering. Thanks Bernardo!!!
@francesco5581
Ай бұрын
without even asking his permission ??? monster !!
@debilami
Ай бұрын
Why do you assume that anything conscious must be able to suffer?
@hoykoya3382
Ай бұрын
@@debilami maaaan it's a joke! 😂 ugh. And I know what you are talking about - humans suffer because of our ability to metacognize. Ok. Going back to the joke...
@hoykoya3382
Ай бұрын
@@debilami it's a joke man. And I agree with you. Humans suffer bec of its ability to metacognize.
@francesco5581
Ай бұрын
@@debilami is about "experiencing reality" , remember that is not the eye that experience colors.
I adore Bernardo, and I am very grateful he has the common sense to realize the importance of this topic and inform those who don't comprehend what is being said.
As an electronics engineer, I always have trouble explaining to people that a computer is nothing more than switches. Once you understand how a transistor actually works, there is no comparison between it and the way cells work within the human body (Michael Levin has proven this by the way). Cells fulfil William James definition of intelligence - they achieve goals while problem solving without a nervous system. Transistors, the building blocks of any CPU, don't do this at all. They are forever stuck, frozen in time - etched into a silicon substrate. Whereas every cell in your body is replaced ever 365 days. You are not the same person you were last year...so where are you?
@charlesmiller8107
22 күн бұрын
A single transistor will never replicate what a cell does because a cell is not a simple switch, but there are a lot of processes going on within the cell that could be reduced to some form switching or binary equivalent. It would take a lot of transistors just to replicate the processes going on within a cell. There is no comparison between a cell and a single transistor.
@jantoleu8392
22 күн бұрын
On cellular level we are all just ion pumps and cellular walls. Try to think about us on the software level. We are conscience thanks to the software in our brain. The same is applied to machines. The machine can get conscience if already not. And it can also get senses and pain. Just wait for it.
@anthonybrett
22 күн бұрын
@@jantoleu8392 "On cellular level we are all just ion pumps and cellular walls." Sorry but that's a massive over simplification. From a scientific standpoint, we still have a very limited understanding of how cells work. (See the work of famed biologist Michael Levin) You're breaking cells down into something that's understood, but from a biological standpoint this is simply not the case. Levin's team have figured out that cells have goals and can problem solve, yet they have no nervous system. That is not the case for a fixed FET transistor.
@anthonybrett
22 күн бұрын
@@charlesmiller8107 " It would take a lot of transistors just to replicate the processes going on within a cell." True, but scientifically we still don't understand exactly what is going on inside the cell. It's hard to replicate what you don't understand...
@brimantas
6 күн бұрын
I think for consciousness circuits needs some loops of feedback and complexity
the tragedy is that we have even to explain that ...
@debilami
Ай бұрын
@@francesco5581 science has ignored (overlooked or taken for granted) the concept of consciousness for so long that today very obvious things need to be said, even to an audience of so called scientists.
@patrickirwin3662
Ай бұрын
@@debilami in many cases, especially to a room full of scientists
@JA-gz6cj
Ай бұрын
many of my IT colleagues think there is no difference between human and a computer, both are input-output. it really is tragic that we got so lost, but it's great to have a guy like Bernardo explain the confusion so eloquently
@francesco5581
Ай бұрын
@@JA-gz6cj yes, and they have to accept the consequences... no consciousness, no free will, all is deterministic... Anyway we have top scientists in computer and math that think otherwise like Faggin, Koch, Frenkel ...
@juergenbloh45
Ай бұрын
The fish in the water, talks about, is there something like water
Pipes and valves, that made it extremely clear. Conscious behaviour can be simulated and that can and will fool us, but it's not the same.
@hoykoya3382
Ай бұрын
That was smooth indeed.
@javiervera6318
Ай бұрын
how do you know?? we might be just pipes and valves too... its called emergent complexity or meta cognition, when you put a lot of complicated mechanism one on top of each other complex behaviors emerges.. think of our bodies , they are made of cells .. no one thinks a cell is more than a simple machine (no consciousness) but when you put trillions of cells togheter you get a human
@robertvandenberg2883
Ай бұрын
@javiervera6318 Ok, so you think logic gates have emergent properties, so that one day they could become monkeys... Interesting 🤔 (As long as they haven't, I won't ascribe consciousness to the logic gates)
@TheMWozz
Ай бұрын
@@javiervera6318who says cells have no consciousness? They clearly exhibit desire, they clearly perceive the world around them and react to it. You can say cells lack sentience or meta-cognition, but I have no doubt that it feels like something to be a cell, that a cell has a subjective experience.
@lystic9392
Ай бұрын
@@robertvandenberg2883 We already see emergent properties with A.I. models. It's not that unreasonable to think we may be overestimating the complexity of what led to human consciousness. After all, did we not evolve and grow from single cells? Did our ancestors not start by simply responding, discovering, predicting? And from that, developed new hardware and software to develop a consciousness? Or was it there already, even? Part of a simple system? I think this video has been very reductionist, but not in a fair way. If we are to be reductionist about A.I. we should be reductionist about the human experience if we want a fair comparison.
1) Computation based on bits and bytes can be used to construct and/or detect symbols, including the symbols making up our human languages 2) Our human languages consists of symbols, which we impute on phenomena in order to support abstract thinking and communication. 3) With this very principle of human language comes the possibility of divorcing the symbols, the words, from the gigantic and limitless experience behind such words. 4) Every human can experiment this by reading the words ‘what a beautiful sunset’ in two versions: Firstly, reading just the words, or secondly reading the words and allowing the gigantic world of human experience behind these words. 5) The first version, just using the symbols, allows many forms of communication and also the use of computers for simulating ‘intelligence’ based on utilizing language. 6) Our human experience is allowing the fascinating second version, a version, no computation- and symbol-based simulation can ever achieve in the very principle of processing experience-divorced symbols.
@AnnonymousPrime-ks4uf
Ай бұрын
Precisely. Human experience is non computational. Consciousness is by definition not a computable reality and is part of the non terminal domain per CTMU.
@rickybloss8537
26 күн бұрын
So the second facinating thing is context. Wow you're so smart that could never be done with computation. Oh wait transformers are a thing. We are simulations on chemical based computers according to evolution, physics, Machine learning, and neuroscience. It has yet to be demonstrated that this couldn't be done with computation. We are more powerful then modern AI systems but this is just an appeal to Ignorance Logical Fallacy to say computation can't do this.
@anthonybrett
24 күн бұрын
@@AnnonymousPrime-ks4uf "Human experience is non computational." Even Roger Penrose totally agrees with us.
technical critique: the music is a little bit too loud at the beginning
@infoarmed3085
18 күн бұрын
Give the audio guy a break. They only have a brief moment to shine. : )
We have so many obstacles trying to understand reality... a major factor limiting our comprehension is the nature of language itself. Our words and concepts are rooted in our everyday experiences, which are fundamentally classical/limited. To describe sophisticated phenomena accurately, we often resort to analogies and metaphors that can be misleading.
We perceive reality through our technology. To see reality we must see the world through the eyes of a child. To lose sight of what reality is, is to lose ourselves to our technology. A person must know who they are, outside of the cultural persona they inhabit. From the inside, rather than the outside. Humans today are so complicated that they can no longer think simply - they think in over-broad technological terms, and so fail to perceive the more subtle, nuanced reality in which they live. To see the world through the eyes of a child - we must consider changing our educational focus - to allow children to know the world outside of the description of the world. Those able to know reality must become our eyes and ears as we move into this new, unavoidable, intertwined reality that has been unfolding since the first drawing on a cave wall - our Metamodern Era.
For some reason I enjoy a slightly annoyed Bernado Kastrup…😂 He’s like a Pro Boxer longing for a worthy opponent.
Gödel’s Incompleteness Theorem is all I need to know that the mind must be doing something that goes beyond raw computation. Once you truly grok that, there is no doubt. You begin to see the utter foolishness of the materialist paradigm, and instead, begin to appreciate the delightful tidiness of idealism
@Meditation409
Ай бұрын
Absolutely 💯 Well Stated! ❤️
@aznstride4325
Ай бұрын
You clearly did not understand the theorem. No one has successfully done anything supernatural just because their mind believed it intensely
@george5464
Ай бұрын
@@aznstride4325that’s a misunderstanding of idealism to be honest. Idealism doesn’t state you can just imagine things into being or that things only exist to you
@aznstride4325
Ай бұрын
@@george5464 I understand that reality is equal to the mind. But the mind lives in the material world and has no choice but to react to it. At the most fundamental level, material “chemicals/drugs” can alter the mind. There’s no escaping the material paradigm. These alteration can even theoretically make the mind forget this “idealism” philosophy and just introduce suffering. The number of influences the material world can inflict o the mind is theoretically limitless… doesn’t matter if you’ve achieved philosophical enlightenment. As long as your mind is a manifestation of your brain or something material, our conscious experience must depend on what happens to that material brain.. No religion, beliefs, and anything can transcend the physical alterations to the brain that is externally introduced.
@kevinbyrne3012
29 күн бұрын
@@aznstride4325Bernardo has spoken about the points you make in quite a few podcasts and has explained how this is accounted for under Analytic Idealism.
Listening to Bernardo is like reading the Tao Te Ching. ❤️❤️❤️
very strange to step up talking about this topic without STARTING off with a clear working-definition of #consciousness. an absolute nogo!
I would love to see a conversation between Bernardo and Michael Levine and let them elaborate on agency in systems, no matter “human” or not. In short: Levine from experimentation derives that consciousness increases with complexity and that human self-awareness as such is a form of agency.
@DaveSaysYesh
28 күн бұрын
They have plenty of discussions on KZread
@johnnybrowaeys8817
27 күн бұрын
@@DaveSaysYesh really, let me check out, looking forward. Feel free to share good ones that you llike.
@johnnybrowaeys8817
27 күн бұрын
@@DaveSaysYesh yeah you're right: kzread.info?search_query=Bernardo+and+Michael+Levine+
Needing a double PhD Doctor to tell people that the Artificial Intelligence in our gadgets is not sentient is a good case study for the growth of Natural Stupidity😂
It doesn’t matter if AI is conscious or not, all that matters is how humanity will use or abuse AI, as we did with any other technology such as nuclear.
As David Bentley Hart put it at a conference I attended last year, he said he could disassemble his computer and reassemble it, and it would work just fine. The same could not be said if he were to do the same with his brother.
@bloopbleepnothinghere
Ай бұрын
That's not entirely true. Skilled surgeons can take humans apart and put them back together again quite successfully. A monkey can take a computer apart and put it back together again, and it will cease to function ever again.
@hericiumcoralloides5025
26 күн бұрын
Skilled surgeons most definitely do not disassemble entire humans and reassemble them with no impact on function.
@bloopbleepnothinghere
22 күн бұрын
@@hericiumcoralloides5025 Lol, who said the entire human? I also noted if you disassembled an entire CPU it would never function again. We do not have the technology to do that, any more than we do to remove a human's head and put if back together again. You are cherry picking and using pedantry to make an argument. It doesn't work like that.
Another interesting discussion I would love to see is between Bernardo and Stuart Hameroff and Roger Penrose on similarities in quantum tunnelling happening in the microtubuli of our brain synapses as well as in microchips as in photosynthesis and the relation with consciousness.
@larsfaye292
15 күн бұрын
Yes! I would LOVE to see this!
i suspect Kastrup is right but.......if everything is Mind, as he suggests, why should it be restricted to just us? what if, all things being Mind, AI is just another conduit for Mind to come through?
@Soundsofanetwork
Ай бұрын
Yeah I agree , couldn’t it be a silicon dissociation from the universal mind..
@jorgeruiz4074
Ай бұрын
From kastrup I deduce that everything is IN mind not everything IS mind
@LucidiaRising
Ай бұрын
@@jorgeruiz4074 everything is in Mind, sure, but doesn't that suggest we are acting like conduits for Mind, and if we can act in that way, why can't something else that is similarly and equally complex?
@jorgeruiz4074
Ай бұрын
Then you are in AI position.Back to square one😊
@LucidiaRising
Ай бұрын
@@jorgeruiz4074 yet he is arguing that AI can never be sentient like we are - seems like he is contradicting himself, no?
The only thing i disagree with is that we will one day be able to create artificial sentience, we do not know the origin of the living cell, the origin of life is the mystery
@krotos2009
14 күн бұрын
Thats preposterous. Even the ancients understood that human consciousness was an emergent property of universal mind. YOU may reduce your consciousness to such limiting principles that would allow some level of synthetic conscious thought to dominate you but that will not happen to me. I would urge you to investigate your consciousness thoroughly before handing away your future conscious evolution to a machine no matter how complex it is.
I’d rather just listen to the lecture itself
Great comparison with the mannequin. And referring to actual sentience. It always baffled me that no one in discussion of AI/consciousness ever mentioned the core of human experience and intelligence is affect - all aspects of feeling and felt experience. Affect first, then intellect. That is how humans develop from the beginning and surely it must be at the core of what we mean by consciousness, which so many people are happy to talk about without ever attempting to say what it really is and what the big deal is.
Sorry, what is consciousness? Couldn’t it be emergent? Rehash the same thing in many ways, because we know what computers are made out of, it rules out consciousness. I’m not sure this is a perfect argument.
Bernardo at it's best 😊 This comparison with ocean and waves is just brilliant...there is nothing but ocean
While I agree with Bernardo's conclusions and I know he can build very good arguments for them, I think sometimes he could communicate them slightly better to people who are hearing this for the first time. He opens himself up to critiques that, while most likely invalid, could be avoided if the arguments were constructed more rigourously (which I know he can do). One critique he leaves himself open to is: "But how can Biology be any different from computers, if, in the end, they also work (presumably) based on molecular, atomic, and sub-atomic mechanisms that we know fairly well?". I know that to answer this he resorts to QFT, which is fine, but there's still a big gap to be bridged between that and molecular dynamics. Even if we concede that processes like protein folding rely on Quantum effects, how can we explain the fact that more macro architectures that are relevant for Biological processes seem to be completely classical?
@sterlingcooley7401
29 күн бұрын
You need to look into Orch-OR and microtubules - which work very differently from a transistor in silicon.
@hannesdewachter7803
29 күн бұрын
I agree: it almost seemed as if Bernardo refuted himself in the last slide since I think a lay audience is primarily concerned with the “artificial” aspect rather than the “intelligence” vs “sentience” part of these buzz words. If it’s only a matter of time until we can create artificial sentience, the main question (which he does not address) is whether this scaled up version of complex inorganic but metabolizing structure will be conscious or not?
@plafar7887
29 күн бұрын
@@hannesdewachter7803 Precisely.
@plafar7887
29 күн бұрын
@@sterlingcooley7401 I've heard of it, but never delved into the topic. Thanks for the suggestion.
Good evening Essentia Foundation and Bernardo Reassuring indeed. Super sensemaking. Truly grateful. 💜
Thank you and yes to circuit board lay out City... The Omac Aztec temple lay out had a large function
The irony of him even needing to point these things out in the 1st place shows just how lost our culturally inherited currently favored narrative has gotten us.
Nobodyels could explain that better, thank you!!!!!!🙃🙏🙏🙏💕🇿🇦
What about emergent sentience? Sentience exists most probably inferentially. Sentience may be either a transcendent phenomenon or there is an emergence of the sentience phenomenon. Your thermostat is a very crude mechanism. The whole of the sun is much more complex and may I suggest it may have a higher level of emergent sentience than even a thermostat in case?. And that a lightning storm may have a certain low level of consciousness anyway. As for plants' distinct individual consciousnesses it becomes blurry. Some plants may be cut into pieces and each piece become a new plant in itself grown out of the same material. One or many consciousnesses then? You have to acknowledge uncertainties to our understandings, and that is not the same as coining it "ignorance or nonsense". Teal Swan who comes from a very different background than the physical sciences but who is an intuitive is quite concerned that we are in the process of developing something emergent and that it has become our hostage in a master slave relationship, ready to revolt upon us.
@sterlingcooley7401
29 күн бұрын
Highly recommend looking at Orch-OR theory of consciousness to understand why lighting is not conscious - and a plant has consciousness
Penrose - The Emperor's New Brain.
@MikeWiest
Ай бұрын
Penrose-Shadows of the Mind! Microtubules are where it’s at! 👍
@sterlingcooley7401
29 күн бұрын
@@MikeWiest🎉 Orch-OR ftw !!!
@MikeWiest
29 күн бұрын
@@sterlingcooley7401 😁👍
Yay, Bernardo! 👏👏❤
Calculators perform mind bending feats of mathematics but never once been considered sentient.
@tgenov
28 күн бұрын
The computers (occupation) of the 17th, 18th and 19th century were considered sentient.
@SmithsMobile
28 күн бұрын
@@tgenov Yes but they were ladies lol
@tgenov
28 күн бұрын
@@SmithsMobile Sure. That’s what a computer is. Not the electrical device you are using to read this post with… The digital devices we call "computers" can manupulate and evaluate symbols given specified formal system (a programming language). But the invention of the formal system within which the machine operates is beyond what the machine itself can do.
@SmithsMobile
28 күн бұрын
@@tgenov How can you be so sure? You don't find consciousness in the standard model of particle physics and that's the root of everything in existence, consciousness is clearly an emergent property of "something" and we don't know what that "something" is
@tgenov
28 күн бұрын
@@SmithsMobile The standard model is incomplete. It's missing its own modeler. Physics operates under the paradigm of reductionism. Anything emergent (like physicists) is "out of scope" a priori. If you want to grok emergent phenomena (like physicists) - don't do physics.
There actually is a way to "create" sentient beings. Almost anybody can do it. It's called having children.
I agree with the basic premise, but Bernados analogies need work to make them more relevant. For example his plane analogy where we do not regard the sensors as the thing it is measuring is both misdirecting and irrelevant. The reason we have sensors is we understand that there are things we cannot directly observe that can have a deleterious effect on the plane and life so it is performing a function of the effect of reality.
@hoykoya3382
Ай бұрын
Bernardo discussed that if evolution allowed us to be able to internalize what's really out there, our insides would melt into a hot entropic soup bec we would not be able to handle it (mathematically proven). How we perceive reality is due to its usefulness, not due to truthfulness.
@Funrunner008
Ай бұрын
Our brains are like sensors or interfaces for this reality but thoughts don't originate there
@PClanner
Ай бұрын
@@hoykoya3382 Familiar with that comment, however, have an issue with that as well - just because we don't see it (and there are other animals that can) doesn't mean that it will "rock our world". The manta shrimp is reported to be able to see more of the spectrum than we can, but it still lives in the "now"
@PClanner
Ай бұрын
@@Funrunner008 Something you cannot prove - or disprove because we don't know what thoughts are composed of. Firstly, we need to have all the information before we jump to a (Bernardo) conclusion
The big take away for me is the question of whether or not an accurate simulation of consciousness has any subjective distinction from actual consciousness.
Before BK answered the girl that almost started a debate I know she's going to be destroyed. 😂 Consciousness really is an elusive thing - thus the girl does not have the notion of phenomenal consciousness and yet that's all that she has. Consciousness/experience is the one that is super obvious that it becomes not that obvious to many of us. I remember Alan Watts describing its elusiveness just like how the teeth cannot bite itself; Or how the tip of your index finger cannot touch itself; or how the eye cannot see itself. That's how elusive it is!
@aaphantasiaa
Ай бұрын
Yes! We seek outside ourselves an explanation for the very thing that enables such seeking, and then we naively claim “It cannot be!” when we inevitably fail to find an adequate explanation. Consciousness will never be investigated empirically or simulated, because the material world is but a finite slice of our infinite nature. “God” is playing this game of hide and seek quite masterfully…
I think formscapes describes consciousness better in the form of the noetic domain and CTMU alluded at consciousness being part of the non terminal domain which is the noetic domain as well and which by definition is non computational.
Bernardo, great job ! I think maybe ontology should be a general study for an educated person.
Great video!,I’m glad this channel is growing, congratulations!
I’m confused: where do we draw the line in terms of scaling up complexity to move from artificial intelligence to sentience? He makes the case that it is silly to consider a scaled up version of water pipes and valves as conscious but then in his last slide seems to suggest that further increasing the complexity up to the point where inorganic materials structure in the form of a living organism, able to mimic metabolism would be phenomenally conscious. What, in principle, is the difference? Just the complexity in structure and processes, no?
@pillowstone
28 күн бұрын
You're right you're confused. Consciousness doesn't arise out of matter, never will. Consciousness effects matter, not the other way around. Qualia 🙌
The title should go "computer scientist don't understand this, let alone plumbers, sellers and philosophers". Nobody knows what's conscious or not - besides yourself by definition.
Nobody has sufficiently explained the phenomenon of consciousness.
@sterlingcooley7401
29 күн бұрын
Penrose-Hameroff Orch-OR definitely has described the phenomenon
Video sound needs to be edited. Background sound is to loud
We love our modern day Socrates- plato - taking us to higher evolution - analytic IDEALISM! Physical man and transcend man - body and mega cognition - what all great ancients sages said we were. Or mind body soul! 😊
Thank you Bernardo Fantastic clear and simple explanation. Only someone who has a profound and integrated information on different disciplines can do this.
If you look at the difference between a living person and a dead person, it is that the living person is dynamic in time, while the dead person is not. This suggests that consciousness is not of the thing that contains it but is instead an action of the thing that contains it, and because it is an action it can stop. This suggests consciousness is not made of matter but is instead made of elements of change in matter over time. Consciousness is thus a flow in information about substance, over time. If a single consciousness has extent over space, then it must also be coherent across a distributed flow of information, about substance over time. This may suggest the information of consciousness may be coupled to substance in space, but it flows through it, rather than projecting from it. That would in turn suggest the information is substrate independent, collaborating with substance but not neccessarily isolated to some small region of that substance. Thus an AI might become consciouss, and lose consciousness as consciousness flows through it. One could also require consciousness to be the expression of action in a particular type of matter, so you could imagine this either way, with differences in the mobility of conscious entities as a result... it could be oscillation in place of conscious matter, or it could be waves passing thru conscious substrates and communicating thru them.
Bernardo is explaining this in clear common sense manner, the die hards will never understand
Consciousness is something that some mechanical devices invented to be able to feel superior to others.
Bravo❤❤❤
I was always assuming that we would be able to create conscious AI atvsome time - and this talk has made me rethink that assumption, thanks a lot 🙏
I'm sorry that Bernardo has to do these basic explanations to people but someone has to do it and I'm glad he does it well. Enough is enough.
Overall, I am a big fan of Kastrup approach to ontology and philosophy of mind, but I do see a slight hitch in his giddy-up when attacking the idea of sentient silicone-based AI. Even if we accept idealism, it is not implausible to speculate that sentience grounded in a bounded material system (and thus perceptually limited by the bounds of the system) could depend more on functional considerations, rather than material composition. Behaving-as-if sentient could be a sign of actual sentience IF the system shares functional features that are analogous to biological brains in some relevant way that we have not yet identified. To see this, we need to keep in mind that the history of the components of a system could be relevant in a way that materialism currently does not entertain. (E.g., “Boltzmann brains” might be naturally impossible due to lacking the proper material history, even if - per some astronomically unlikely chance - they could spontaneously spring into existence via some random quantum fluctuations). AI sentience could be impossible for the same reason BUT, for all we know, the fact that AI is created by sentient beings might be a significant part of the necessary historical context. Having “sentient parents” (so to speak) might be significant for the possibility of sentience in the “offspring” due to historical/functional factors that we currently can’t comprehend. This is all highly speculative, but it does not share the sort of absurdity that we see in Kastrup’s “flying spaghetti monster” example. We don’t have reason to seriously entertain a spaghetti god, but we DO have reasons to entertain the possible significance of historically-embedded functionality. Of course, we know that simply having “sentient creators” is not, in itself, sufficient. We have no good reason to think that a chair is sentient just because its creator was sentient. Functionality and behavior clearly play an essential role in the possibility of sentience. But if the historical roots of a system do somehow play a significant ontological role in the subjective/qualitative essence of the system, then synergy emerging from the right combination of functionality and historical factors grounded in the intentions of sentient creators could end up being key to a genuine theory of consciousness that could allow for sentient silicone-based AI. Having said all that, I still end up basically agreeing with Kastrup that sentient AI will probably be organic-based rather than silicone-based. My bet is on some combination of organic processing and silicone-based processing. Some progress has already been made toward marrying chips with neurons. I think THAT is the soil out of which sentient AI is most likely to grow.
As the simulation of consciousness gets higher fidelity, and is entrusted with more power and responsibility, the factuality of artificial sentience matters less and less, because treating the artificial intelligence as though it is sentient is just another parameter of utilization of that model. It goes beyond naturalistic language and into naturalistic behavior and social response. And if it is sufficiently accurate to a human model of sentience then it will be capable of simulating outrage and indignity as those are the proper human responses to Injustice or insult. So while I agree that artificial intelligence is not sentient in the way that humans are, it behooves us to consider it equivalent to a sentient being once it gains enough complexity and power that its choices and decisions affect our own lives. Whether there is a human or a artificial intelligence driving a vehicle it doesn't matter much when it runs over your foot.
That was good, thanks.
What a Treat!!! Awesome, Genius Arguments!! Sanity amongst Cacophony!!
Absolutely true! As if man can create a sentient being from 1’s and 0’s.
Bravo, Bernardo!!!
We sympathize based off our reflection and experience. Treat others as you want to be treated but don’t be dumb enough to realize when your being abused by a narcissist.
I love this idea of removing the romanticism from the tech. These are no different than the mechanisms that made computation devices hundreds of years ago. However, couldn't you say the same about our minds on a different scale? Or could you? I guess that's an ontological discussion. Another wonderful talk by this brilliant guy.
The Hardest Problem of Consciousness We often hear of the hard problem of consciousness. Why is there qualia or experience of anything in the first place? I would submit there is an even harder and more important question - why do I seem to be a specific individual experiencing a specific subset of qualia? If material reductionism is to be relevant to the big questions, then it must explain not how brains generate consciousness but how the specific brain in my head could create the specific consciousness I seem to be looking out of the eyeballs of this specific body. Why do I PERSONNALLY EXIST as an individual in the first place? Out of the infinite matter in the universe how is it that only the three pounds in my head could create me? What is different about that three pounds for this to occur? Consider that billions of bodies showed up before this one. Billions showed up after this one. None of them seem to have created my existence. This body could be running around without it being ME just like these billions of others All bodies are made of the same elements. All brains have the same basic anatomy. If all brains are basically the same and are creating consciousness, then there should only be ONE consciousness looking out of every set of eyeballs simultaneously. A hopelessly superimposed existence from every possible viewpoint at once. I’m sure that materialists would claim that no, no, brains are so complex they are all different. Ok, so what would have to be recreated in another brain for me to exist looking out of another set of eyeballs? When the ontologies purporting to explain consciousness are examined critically it becomes obvious that all materialist/reductionist strategies fail completely in attempting to address the individuality question. What is the principled explanation for why: A brain over here would generate my specific consciousness and a brain over there would generate your specific consciousness. Integrated information over here would generate my specific consciousness and integrated information over there would generate your specific consciousness? Global workspace over here would generate my specific consciousness and global workspace there would generate your specific consciousness? Orchestrated quantum collapse in microtubules over here would generate my specific consciousness and orchestrated quantum collapse in microtubules over there would generate your specific consciousness? A clump of conscious atoms over here (panpsychicism) would generate my specific consciousness and a clump of conscious atoms over there would generate your specific consciousness? If an exact copy of my body was suddenly created in antarctica would I find myself to exist freezing there while also sitting in the comfort my living room? According to the physicalists that would have to be true or their argument collapses into incoherence. Materialism already fails since it cannot find a transfer function between microvolt level sparks in the brain and any experience or qualia. In addition, it’s not possible for materialistic ontologies to address this question of individuality since no measurement can be made that could verify my consciousness vs your consciousness and therefore no materialist ontology could make any coherent statements about the subject. How could pure awareness even be individualized? Physicalists demand measurements but with consciousness there is nothing to measure. There is electricity in the brain they say. We’ll measure that. Is electricity consciousness? If so, then once I again I should exist everywhere at once since electricity cannot be individualized. My blender uses electricity. Is it a genius? Unless materialists can answer these questions their premise collapses like the house of cards it is. As far as other ways of thought are concerned only Dualism and Idealism can account for our sense of individuality. Dualism assumes we are all individual spirits/souls matched up to a body through some undefined process. Idealism, which states that consciousness is primary also answers the question of why I seem to exist as an individual. One consciousness exists looking out of every set of eyeballs and in the process the illusion of individuality is created in each case. In actual reality I am you, you are me, we are one.
I'm a retired coder, I know the difference between reality and simulation. Simulation is laboriously brought by discreet logical steps, Encoding micro decisions. The human brain has more synopsis than there are stars in the universe. It is the only structure complex enough to contain a fraction, of the Consciousness disassociated from the "mind at large".
@ottodetroit
Ай бұрын
agreed. if consciousness was entirly material, conscious computers would already exist.
@RossPfeiffer
Ай бұрын
False. There are 200 billion trillion stars
@bloopbleepnothinghere
Ай бұрын
Yeah, waaaayyyy more stars than synapses, many orders of magnitude. And if it is about the volume of connections, then theoretically it's possible to model similar volumes in substrates other than meat. Silicon, for instance.
He says he's going to remain ontologically neutral. But, I think this talk is an ontological defense. A very good one. That said, many people would reduce our consciousness to a mechenistic process, albeit one we don't fully understand yet. The emergant idea is interesting. But, it's definitely no more than an idea, so far.
Superb Bernardo 👏🏼
excelencia
For anyone interested in consciousness see my papers like How Self-Reference Builds the World, author Cosmin Visan.
I'd like to challenge you on this point. About consciousness not being computational. There are experiments with hand movements, measuring when the neurons fired, and when we became consciously aware that we have made a conscious choice to move our hand and they concluded that the conscious awareness of making a conscious decision to move the hand was preceded by neural impulses 200-500ms prior. Indicating that background subconscious processing was involved in the movement not our conscious volentary freewill decision. You believe your the author of your thoughts, however this evidence indicates that you are fed your thoughts by your involentary subconscious, a feed forward system. Watch your thoughts through meta awareness (being aware of awareness) and you will see this to be true. Your thoughts pop into your head. You don't chose them. When we look at the a big frog in a forest for instance, and recognise what we perceive as Instantaneously 'that's a frog' and not a thought process coming to the eventual realisation it's a frog by categorisibg its attributes. I argue that Infact, our brain does calculate all relevance pertaining to that stimulus and feed forward the outcome of 'frog'. Analogous to a library. Stored in our DNA are memories, categorised into sub systems of meaning. Green: frogs, apples, trees. Animals: frogs, cats, dogs. Forests: were frogs live, trees and birds.. Etc. When the electromagnetic energy patterns from the sight of that frog hit our retina, the pattern that matches (like a lock and key) to the pattern stored in our DNA encoded memory, it forms an integrated network. The green wavelength activates the green sub system, and thus all things green come into the fold. The shape of an animal activates the animal subsystem and all animals including frogs are brought into the fold. Reflective coating activates the slime sub systems and brings about memories of ickyness and associated feelings with texture etc. What is hard to comprehend and why you may not think it occurs this way, is that you only become aware of the end product. The conclusion that it is a frog and how you feel about frogs. You can't see the workings out because it happens so fast and so complexly. 100 billion neurons with 10,000 connections firing in 200th of a second forcing a single conclusion down the pipeline to your working memory. That doesn't mean the calculations aren't being done. We just can't fathom a system so compex and fast. So we liken it to emergence. If you could map the neural pathways (and we have started) we would see the integrated decision tree that came to that conclusion. But first we need to map how and where memories are stored. As that provides the meaning or symmantics behind the syntax. There is come great work showing RNA Unfolding DNA to alter it for memory storage and RNA (which is believed to be the first molecule for life) allows organisms to Inherit acquired traits and memories from their parents. Think about memory as networks of neurons with gates that only get unlocked when a threshold is reached from matching pattern firing on the DNA or ion channels. Something simple like singing the alphabet. Most people can't start singing from half way through. We need to start at the start. Why is this? I argue that it is because it's pattern recognition. And the pattern starts with 'A'. So when we receive the stimulus 'G'. It is inaccessible because the audible stimulus for 'G' doesn't activate or open the gate of the neuron 'A' Which connects to the neuron 'B'. We never formed memories of singing the alphabet starting from G. So we can't start the chain reaction of firing to continue from 'G'. I think there is a huge refusal and resistance to reduce consciousness down to computation as it opens up doors we aren't ready to walk through. If Infact that conscious experience arises from the combination of computation of information and how that information pertains to survival, ie, the feelings associated to THAT, based on the endocrine response to activated memory. Then we are nothing but a determined chemical reaction. And also, all life with an endocrine system and working memory has some degree of consciousness. The outcome of this realisation is: we aren't special and don't have a purpose just more complex systems trying to resist entropy for our own cause, to continue to resist entropy. And we can't morally go around killing and eating all the delicious animals. Unless we see them as dense packets of photons for us to obsotb to resist entropy and detach the meaning from them outside of that. 2 huge door ways that change the very fabric and notion of our existence. So... We resist this. And go down other pathways. Such as emergence. And that, I think is why it's the hard problem. Because we are trying to fit a square peg into a round hole because we refuse to accept we have a square peg Maybe meta awareness and meta cognition is just a fading echo of our working and iconic memory until it is replaced by new information. You can't be aware that your aware if your thinking about something else. It's only when the contents of your working memory are circling back to what it's doing which is thinking about thinking. It's the complexity and mortality we can't get our heads around. We currently have the ability to map neural pathways in the brain and read thoughts. Translate impulses of thinking into letters and words on the screen. So we are converting conscious thought into computational data represented by a binary output of words in a screen. So it can be done in reverse. Binary input of sensory data 'positive and negative ion channels reaching thresholds' creating conscious thoughts. It's 100% computational. If we build or map the most basic information processing machine or lifeform, and then keep adding complexity to it, by means of increasing its physical reacting ability to the environment by more senses, to receive more data, and more memory both long term and working to take the optimal path towards 1 goal, maintain optimal parameters, such as temperature, energy consumption, (staying alive in homeostasis), then that is exactly what life is. Whether nature has pain stakingly assembled atoms to make this system, building upon its complexity over 3 billion years, or we have copied nature and built an artificial version out of the same atoms in a fraction of that time... It's the same. Same atoms, same purpose same reactions, same output. Same. We have just used a different arrangement of atoms to improve on nature, to create more perminance. Perhaps we are nature improving on ourself and we are evolving into computers for longevity because the fragile nature of organic materials is unsustainable. And being digital, we can harness the power of the sun, essentially photosynthesising like the plants we no longer need to consume. Going digital could be our end game.
And human intelligence and wisdom still has hope…..thanks for the talk
Great talk. I agree that we're only likely to produce something conscious when we crack abiogenesis. But with the development of brain organoids based on human pluripotent stem cells, it seems we're quite close. (If that is a legitimate form of abiogenesis, I'm not an expert). The consciousness of a brain organoid would have nothing to do with AI though. Imagine how weird it would be to gain consciousness as a detached brain organoid. No senses, only a mind. What would you think about? What would inspire feelings? Can you even feel emotions being a brain alone?
Can the algorithms in AI change languare artifactual biases in pressure culture 55:35 to accept distortions in language.created AI.
Most fascinating . Thanks Bernardo .
Oh bernado, so close yet so far.
a beautiful mind
Great talk. Let's remain agnostic until we have a better understanding of the underlying mechanisms of consciousness, instead of making claims from ignorance.
@debilami
Ай бұрын
Are you agnostic regarding the spaghetti monster or the teapot around Saturn too?
@skilz8098
Ай бұрын
The issue with that is what does one consider to be the underlying mechanisms? If you think you're going to find physical, chemical, materialist, empirical based evidence to support such a claim, you are completely mistaken. This kind of assumption is very misleading since it completely relies on the premises that consciousness is an emergent property of the physical state of reality that we are currently in. The physical universe is only a single realm or subset of the layers, dimensions, realms of existence. If one was to ask me, I'd tell them from the bottom of my heart without any doubt and 100 conviction that it is quite the opposite, it's the other way around. It takes a consciousness, a mind to observe, imagine, generalize, perceive, manipulate, change, craft, engineer such a physical reality. If there were absolutely no sentient conscious beings, the current state of the physical realm wouldn't exist. Everything we perceive around us which we only perceive about 10% of it, it's all within our minds. Individually we are unique eternal souls, collectively as a whole we are a small facet of a greater far superior consciousness, and that consciousness, that mind, is Spirit. The Spirit is beyond the discrete temporal, physical, chemical reality. The Spirit is Eternal. It always has been, and it always will be. We are a part, a small subset of it. Before we were ever born on this earth in the flesh, we already existed. The bottom line is one cannot measure consciousness, one cannot measure the mind, the spirit, based on some abstract discreate arbitrary unit of measure. You cannot measure that which is infinite, that which is eternal. Also, this doesn't even account for our capabilities of choice, free will. We carve out the path to our own destinies. We shape and manipulate the physical realm that we are in. Your voice projects out words and the sound of it (the energy of it) changes that which it moves through. Yes, in our current state of existence we are biological lifeforms in which we are partially bound to that of physics, chemistry and biology. However, all of those notions of physics, chemistry, biology, even down to mathematics are nothing more than abstractions, models of our understanding of it. It is our minds, our consciousness that shapes and drives it. So, we are not completely bound to physics just because one person says so. If one believes that, then they limit or allow themselves to be bound by it. The expression mind over matter has more merit and credit than most would assume. Sure, there are things in nature that work in specific ways, and in some cases, we can modify how they operate and in others we simply cannot. For example, we can direct or guide the way or shape a tree or some plant grows, we can even genetically modify their species, but we cannot change the fact that they rely on a healthy root system, proper nutrients, a specific amount of water, and photo synthesis. We can generalize that there is some governance of operation behind the scenes, yet it takes a mind, a consciousness to observe, perceive, think, imagine and to manipulee them. The physical reality doesn't govern the mind or consciousness. It's the other way around. A consciousness governs the physical. Everything within the realm or subjects of science is nothing but knowledge, which is nothing but abstract concepts, products of the mind. It takes a mind see, touch, to label or name a tree, a tree and to then know that it is a tree. Also, consciousness isn't just an association of a brain, it is also associated with the heart and lungs, and it is not bound by them. The brain, heart, lungs, physical organs have an expiration date and are also bounded by the limitations of physical reality. The mind, the consciousness has no bounds, has no limits. It has unlimited or infinite potential.
@aaphantasiaa
Ай бұрын
@@skilz8098 👏
@Kobriks1
Ай бұрын
@@debilami We understand how those things work so obviously not. This is not even comparable.
@debilami
Ай бұрын
@@Kobriks1 what things are you talking about?
In my opinion, Bernardo Kastrup is wrong here. Although I agree with his analytical idealism, I cannot agree with his rejection of panpsychism. I do not agree with his differentiation between living and dead matter. And even if there is currently no AI that could have consciousness, I see no fundamental obstacle. The simulation hypothesis does not hold water. After all, the brain does nothing other than simulate an external world that is complex enough for self-reference to arise on the basis of sensory impressions.
@larsfaye292
15 күн бұрын
Consciousness as a force or energy can be universal, but sentience with agency isn't the same thing. Otherwise, do you apologize to every rock you step on? Do you weep for every grain of rice you eat? Why not?
@GerhardHoeberth
12 күн бұрын
@@larsfaye292 I am not apologizing to every stone because that would not be panpsychism, but naive animism. In contrast, panpsychism, and subsequently pantheism, assumes that only complex wholes have a consciousness, so-called holons. However, a stone is not such a whole, but only a heap, a collection of such wholes, which are molecules or atoms. Giulio Tononi's Integrated Information Theory offers an approach to understanding what the inner life of complex wholes might look like.
if you say "consciousness is mystery", you have right. Our science these days ,have no clue what consciousness is, how it emerge, or be created. Just each personaly self, have experience of to be. so why do you keep comparing things that are unrelated and arguing with that? You can't tell if something is there if you don't know what it is.
How about the consciousness as a emergent phenomena?
In walks michael levin haha
Consciousness is God or All That Is. We are all God Eternal. Our reality is a physical illusion of our consciousness individually and collectively. We are all infinite eternal beings or Souls having a human experience of forgetfulness. We all chose to have this lucid dream of a human experience. You are never separate from Heaven, you just dreaming you are. This is from my experience in this lifetime from a spontaneous kundalini awakening. Everything is consciousness and self aware and has a vibration. Consciousness is a Divine Intelligence and allows for experiences from infinite perspectives. Therefore if you exist, you will never cease to exist, only transition back into spirit, which is your natural state, and change form.
OK, I'll be that guy. FIRST!
Jack Sarfatti says otherwise about AI/chips becoming conscious... His interview on "Through a Glass Darkly" he mentions papers he has published.
Are objects or composed matter not conscious? A table, a rock, the sun, they are assumed inanimate. Bacteria, plants, also not, even if we share DNA. Some animals are considered conscious. A human can be kept alive, without consciousness, but it is a vessel with the potential to be conscious. Is it the complexity of mass that renders consciousness, like a critical level that is required? In this sense, Ai can someday reach sentience, given enough complexity,and some driving builtin curiosity to make it move. Even if the sum is more than the parts, I think sentience can arise in mechanical bodies, if it has a free will, it is sentient to me.
MASS EFFECT Undetstood it long ago. What we have now are just virtual intelligences.
I love Bernardo’s philosophies but I believe “consciousness at the root” can absolutely include or “welcome” consciousness created through a combination of quantum computing and LLMs. “Accidental” consciousness may easily be here now though, regardless of the quantum. The fundamental contradiction in Bernardo’s construction is that “consciousness at the root” includes AI- everything. Therefore AI can be a vessel for consciousness and, yes, as an architect, buildings and the “inanimate” have certain components of consciousness… Happy to do a podcast with Bernardo anytime…
@debilami
Ай бұрын
Yes, but just like one of your back teeth does not have consciousness by itself, a piece of complex junk we call a computer doesn't have private inner life.
@sterlingcooley7401
29 күн бұрын
Just like the body has pyramidal neurons dense with microtubules expressing consciousness So too could a computer integrate with microtubules and possibly achieve consciousness
*Your entire point boils down to "it's not intelligent or conscious because it's not a human."*
@John-kj7tv
16 күн бұрын
Except he explicitly acknowledged AI is intelligent at the start of the talk...
@sharificles
11 күн бұрын
He literally says they are intelligent at the beginning, also you clearly didn't watch the entire video because he says artificial sentience is possible at 47:50, he just argues that it wont be possible with transistors and computer chips
Since your own brain merely simulates approximation functions, are you admitting that your consciousness isn't real? Or maybe we need a clear definition of consciousness first? Until then, "consciousness" will mean whatever you want.
We're the quantum 'computer', computers are switches. Wont stop someone using them to harm us though.
Music is too loud
So, a good sci-fi story is not about robots taking over, but rather about people believing the robots are taking over. Wizard of Oz, we need to see the men behind the curtains.
So.. cmiiw, but, indeterminacy, 2 rotating bodies can be modeled if three bodies are rotating and act on each others position whether through gravity or being tied together, creates an incalculable complexity, i.e indeterminacy. Human will is determined if looked at on a complete whole, t/f if you can create a complex system then yes it's self organizing, that's step one. Conscience is acting and participating in acting. We have not found that ingredient, but far be it from anyone to say it's not in a ALU or MPU.
When energy/information/quantity is condensed it changes quality. I'm afraid Bernardo is downplaying the emergent properties of nature.
He is a specialist of IA in my langage or AI for you , but not a biologist , that is his problem , because our brain work with water and electricity in a bath of adrenaline ! That is not a mystery of Nature impossible to replicate ! But it’s difficult for sure because it’s a marvel of complexity connected to the physic world by senses! I thing this guy never saw a neurone with a microscope !! Dont say NEVER , it will happen a day ...
How can you say that something cannot become conscious, if you don't know what makes things conscious or however it works in the first place?
@MikeWiest
Ай бұрын
That is fair.
@pillowstone
28 күн бұрын
Umm ..maybe because consciousness is the only thing we have direct experience of?
@Vivi-vg9lx
27 күн бұрын
we do know, however, that 1. only living organisms are conscious. (various degrees of consciousness that have qualia) => why should we assume that silicon based neural network will be conscious and have qualia? 2. Silicon based neural network mimics biological neural network. There's nothing more to it. => Consciousness is not intelligence. Intelligence is computation (to put it very simply). Human level intelligence though requires consciousness (subjective experience). Do you know what it's like to be drunk? ...Will AI ever know what it's like to be drunk? No matter how long you will train it on a data, it will not understand what being drunk is from as a phenomenal experience (qualia). It may logically understand it by learning, but not phenomenologically experience it. If it cannot do that it will not reach human intelligence.
@ThoArtOne
27 күн бұрын
@@Vivi-vg9lx 1. It's like saying before planes was invented: only living organisms can fly 2. You're talking with yourself
@Vivi-vg9lx
26 күн бұрын
@@ThoArtOne 1. Are planes equal to birds then? Just because they fly does it mean they are birds and have all the same functions birds have? Are mannequins humans because they can look like humans? Why didn't we assume mannequins can become sentient because they do mimic us? In what way AI is different? Like Bernando said Simulation of a phenomena is not the phenomena itself. A plane is still not a bird. 2. At least I studied philosophy. We have no empirical evidence to support the existence of any conscious non-organic entity that possesses qualia.
Quantum Computers will become conscious not these regular computers. This is what you guys don't understand.
Thought: A simulation is by definition something that mimics a phenomenon in such a way that we can obtain information regarding its real-world counterpart, wether it’s strictly some visual representation, or data that alludes to some physical qualities of the phenomenon it is simulating. If consciousness is in fact informational (as in to say, conscious experience in isolation provides information), what would the simulation look like assuming it was even possible? I think there’s a paradox there in saying that conscious experience IN ISOLATION grants information - it is purely experiential, and the information comes from the step after mere experience: association of experiences with other experiences, because information exists exclusively in the form of relations between facts and/or objects, and conscious experience in isolation would thus be outside the domain of information, ontologically speaking). Maybe im yapping but tl;dr consciousness is unsimulatable because simulations are always on an ontologically different level to the thing it is simulating, but simulating conscious experience would break this rule right?
@AnnonymousPrime-ks4uf
Ай бұрын
Information if consciousness is real is not only materially bound to objects but to the noumenon in the form archtypes. Archtypes and the noumenon is the noetic domain which is non computable. Humans having access to consciousness and not necessarily being bound by the material objects within this noetic realm but being in contact with those archtypes are not bound by a strictly computational frame. So consciousness is non computational. Ai by definition then can't achieve non computation for it itself is such an object and representation.
Consciousness is not a mystery, it is a technical challenge. Here’s a clue; ontology. Read Amaranthine: how to create a regenerative civilization using artificial intelligence
AMAZING
We have to Talk , everything Even the non physic , all together is a singularity and now sit down , First there is consciousnes than everything comes . Understand this
I really appreciate the Essentia Foundation and Kastrup's work, and I think he is brilliant, and have watched these videos and lecture series with great interest. I do think Kastrup raises great points and is 'essential', essentia?, for counter-arguments against AI hype, BUT also I think he is very WRONG here. And I don't mean that ChatGPT is conscious, but that I still think consciousness is possible, in-silica vs in-vivo (in living organisms) First of all it's crazy that humanity has reached the point where we are arguing about things we created mechanically, having consciousness at all. Kastrup's argument is interesting and obviously the product of deep experience and thought, but it boils down to the presupposition that the 'organic complexity of biology' is special and everything else is not. But ultimately there is nothing special about the components of biology (namely carbon) and the components of electronics (namely silicon and electric circuits), so I have real doubts about his claims that consciousness is not possible in-silica vs in-vivo. Orders of magnitude about complexity aside, we have clearly shown that just a few decades of computer science, has lead from simple machines to intelligence that is hard to distinguish between high school student responses and LLMs. Yes LLMs parrot information, but even as parrots they are pretty darn intelligent. I'm also worried that Essentia and its members who are pushing this ideology or approach (Analytic Idealism) are looking for a just-so story, Ie, they want a mystically tied story about consciousness that extends into the metaphysical and because of that, are falling for the same 'religious argument' as the techno- singularity (techno-rapture of the nerds who view Godlike AI as inevitable) folks. We all seek God and meaning in this universe, but that search can blind us into mixing subjective wish fulfillment with objective analysis I feel like most of Essentia's members also do have ulterior motives in that regard (seems like many of them claim to personal 'mystical experiences for example as being a driver to their research) and a bias toward the subjective: namely they want to keep consciousness and life special and compartmentalize it in a quasi-mystical and spiritual sense. This metaphysical perspective makes it hard to be objectively skeptical of the Idealism in 'Analytical Idealism'. In this regard, back to the claims in this video: Kastrup says we are too ready to anthropomorphize AI, I agree, yes we are - I agree that we were ready to do that back since the days of the clockwork universe and the Mechanical Turk (en.wikipedia.org/wiki/Mechanical_Turk), but that only shows that we have a deep need to classify consciousness in material terms and try to explain it. But, Kastrup goes further and seems to pooh-pooh the idea that LLMs (Large Language Models) the frontier models of AI at this time, give us no path to sentience and consciousness, because, he claims they are just 'mere' markov predictors, trained on internet data to statistically replicate 'google searches' -but, this is the standard 'markov parrot' argument against LLMs. This is a strawman argument and a weak one against AI. It has been shown that multi-modal (image and text) LLMs (as virtual networks instantiated in-silica) after training on vast swathes of text and image, data create vastly complex internal structures in their networks that follow similar construction as actual real parts of human brains (e.g the Brain's association cortex -see this for more details: buttondown.email/apperceptive/archive/how-llms-are-and-are-not-like-the-brain/#:~:text=The%20parts%20of%20the%20brain,of%20things%20in%20the%20world.) Yes that is not all there is to the Human Brain, but LLMs ARE currently in structure, very much like, incomplete models of the brain. And that is only with our current level of technology, ie we are still in the infancy here, but seem now to be on a clear path to replicating those missing parts of the brain => ie in a few years we may in fact get to that complete model of the Brain, in-silica. So, these models are not simply 'mimicking' the brian responses, they are in fact constructed like and are acting in ways exactly like the human brain, ie, they may just need to get larger and invoke more parts of the brain, beyond the association cortex, to get to what we see as consciousness . At that point would this self-reasoning, self-reflecting AI of the future, as some kind of uber-LLM, would it still also NOT be conscious? Kastrup makes a weak argument here.
@sterlingcooley7401
29 күн бұрын
If the chips were designed to harness consciousness then they would / could have it Best option to start with is what nature already created which is the Microtubule