15:17 "the word beautiful to me means I'm in a state where I can't see all the flaws" { wabi-sabi? from wikipedia: 'wabi-sabi represents Japanese aesthetics and a Japanese world view centered on the acceptance of transience and imperfection. The aesthetic is sometimes described as one of beauty that is "imperfect, impermanent, and incomplete.' }
Wow!! I didn't knew Marvin was such a thinker. The knowledge he had was more profound than Ray's. Amazing. Thanks Marvin for all that you've bestowed to the humanity.
I didn't know this interview occurred. Interesting to see Ray Kurzweil take the position of asking the questions, especially when the interviewee is Marvin Minsky.
5 лет назад
Noam Chomsky could have explained how far away from being able to answer this question we are as to render the question pointless and even conceited. There is no such thing as artificial intelligence since we don't know what intelligence is.
"I'm not the same person I was 30 minutes ago, especially after having talked to Ray Kurzweil for 30 minutes". Thats a great complement from one friend to another.
I think Minksy's take on emotions (that any given emotion represents an instance of turning off or suppressing certain critical faculties) is a helpful insight. The pleasure of an emotional response is undeniable, but almost all decisions made in a moment of excessive emotionality, whether the emotion is a "positive" or "negative" one, turn out to be ones we regret. I've tried to train myself to defer decision-making, and sometimes to avoid speaking, in the heat of the moment, but it's much harder than it seems.
Marvin was one of those amazing people. I enjoyed our conversation. I studied linguistics under one of his first grad students. All part of who I became intellectually.
wow, the part where he said "we don't have any trouble putting 50,000 people in a stadium to watch some people kick a ball around" was a really good way to put it. we really will just become creatures of leisure, and that doesn't seem like a bad thing. we already basically are, just look at people who spend their weekend watching tv or playing their xbox all day.
Why? He would trash it as completely nonsensical and would lobby hard to get these GPT services shut down. He was a hardcore symbolist who purposefully made a career out of destroying the work of neural network scholars.
@@oldtwinsna8347 okay, you know nothing about Minsky, the history of AI, neural networks or anything else. “Lobby hard to get these GOT services shut down” on what precedent would he even consider that a possibility. You have absolutely no idea what you’re talking about. 🤣🤡
3:45 "Hundreds of different structures and they each work in different ways". I remember Jeff Hawkins saying the brain works the same way regardless of whether it is processing speech or vision or whatever :- Hierarchical Pattern Recognition.
@@anthead7405 Conserved? Maybe. But preserved, to the point it could be functional again? How will they prevent all the water in the cells from swelling and irreparably damaging the brain once they freeze? Among other problems of course
The threat of a supercomputer becoming conscious and enslaving all of humanity is overblown, in my view. No matter how efficient it becomes at reaching its goals, the underlying “preferences” deciding what goals to strive for would be pre-set by the programmer just as the preferences we have as members of a species is set by our biology. So the only risk of a supercomputer taking over is if it either (1) is used as a mere tool by some human agent, or (2) we are able to construct self-replicating machines that are able to evolve over time according to Darwinian principles and then placed in an environment where survival and reproduction depends on the ability to dominate. Otherwise, similar to an ant, no matter how “intelligent” it becomes, it wouldn’t start to act on other “preferences” than those it comes already equips with. And unless it’s been programmed to dominate or to ensure its own survival, it wouldn’t start acting self-interestedly any more than an Einstein ant would suddenly stop placing the survival of the collective over that of his own. Nor would a computer or an ant, no matter their calculating power and “intelligence”, see any point in changing those pre-set “preferences,” even if they somehow could.
The bigger threat with AI, I think we're seeing now in the present with things like Sora and other image generators, is their ability to be used to distort reality and misinform people, potentially really effectively. People should be focusing on this, because this poses a serious threat to being able to make a judgments, have a conversation, keep abreast of current events, and stay objective.
The chief concern imo is the implementation of it deliberately by nefarious humans with totalitarian intent. Mass surveillance, social credit databases instantly accessible by everpresent facial scanning CCTV, and other horrors perhaps as yet unimaginable-but with humans as I say behind it.
Kurzweil: so if we scanned your brain, we had a really high-resolution scan and the ability to recreate you, you'd be less concerned about skydiving and if something happened to Marvin Misnki number one? Minski: absolutely, but I probably wouldn't go skydiving anyway because my time is too valuable. RIP
Imagine being trapped in this life for all eternity. What a terrifying nightmare. People in power could punish you by locking you in a coffin for 10,000 years. A true hell on earth. And what if there is a after life and this is all just a learning experience. They are denying themselves the chance to find out lifes true purpose.
Consciousness, perception and thinking. A theory of mind according to platonic physics. You will not find an explanation as understandable as this in the current Stanford Leibniz site, which is incomplete as it makes no mention of Mind. 1. Plato's Mind (the One, the Self) is the cause agent, the singular cybernetic control point, of all perception, thinking and doing in the universe, where control is top down from Mind. 2. Plato's Mind is timeless and spaceless, and being the only Reality, time and space are not ultimately real, but are artificial constructions. 3. Since Mind is mental, not physical, all control and causation is mental, not physical, and top down, since Mind is the singular (cybernetic) control point at the top. 4. Thus Mind plays the brain like a violin, not the reverse. 5. Man's mind (small m) is a passive mental subset, or monad, of Mind and under its control. 6. This monad (our mind) is the mental correspondent of the brain and controls it. Our mind plays our brain like a violin. 7. Thinking is the intentional action of Mind (and thus mind) on mental entities such as ideas, manipulating and transforming them intentionally (through will). 8. Qualia are simply sensory experiences, the conversion by Mind of sensory nerve signals into mental sensory experiences in a fashion similar to the conversion of physical sensory nerve signals into mental images. 9.. As Dennett has explained, In materialist thinking, there is no end to homunculi viewing the universe through a chain of homunculi. Leibniz terminates this infinite regress by making the last viewer the Self , which is at a higher level and suitably equipped. 10. Perception occurs as Mind converts physical sensory signals in the brain into mental experiences in one's mind. 11. These experiences can be made conscious (are made aware) by reperceiving or thinking them. This is called apperception by Leibniz. Thus consciousness is apperception. 12. The universe, according to Leibniz, is viewed directly by the One (the Self, the ONLY true perceiver), which views these scenes discretely and in sequence (analogous to snapshots) at discrete points as a whole indirectly through the totality of individual monads, and from their own perspectives. 13. This totality of sets of individual perceptions is then distributed in the proper order and perspective to each of the monads in the universe. 14. These individual sets are called "perceptions", and must be distributied in this indirect fashion by Mind because each monad, in order to remain an individual, has no "windows", to use Leibniz's term. 15. The perceptions are made up of what the monad would see of its nearby neighbors if it were allowed to do so. This is purely mental, but allows us to speak in terms of spacial distances and directions, through these snapshots, between physical bodies, which Mind, being spaceless, cannot actually directly. 16. Mind is also timeless, so that time is physically "created" as an artifact through the actual motions of physical bodies in physical spacetime. 1 17. Intelligence is the nonphysical ability to freely make autonomous choices. It is a faculty of nonphysical Mind, the Nothing out of which the physical universe exploded in the Big Bang. 18. Another name for this nonphysical intelligence is "life." Leibniz maintained that the entire universe is alive. 19. Each monad is perpetual, created at the beginning of the universe and only annihilated by Mind. 20. Since monads can contain other monads, they can. as plants do through seeds, and humans do through sexual reproducxtion, produce subsequent generations. 21. A robot or computer has no Mind or Self which has the wide bandwidth, intelligence and intentionality to actually perceive , think, or do things, such as Mind does. So, being without Mind, computers can have no actual intelligence or life. 22. The current theory of mind is materialist. In contrast to the above, it uses the usual decapitated, mindless, or where mind is at best an abstract entity, not a living presence as in the above. The materialist model of perception, thinking and doing, being Mindless, is dead. DSG Dr. Roger B Clough NIST (retired, 2000). See my Leibniz site: rclough@verizon.academia.edu/RogerClough For personal messages use rclough@verizon.net
"A childish person becomes anxious, Thinking, "Sons are mine! Wealth is mine!" Not even a self is there [to call] one's own. Whence sons? Whence wealth?" Buddha How interesting, isn't it? :)
"described by Isaac Asimov as one of two people more intelligent than himself" lmao, Asimov had a pretty big ego, although I agree with Asimov that Sagan and Minsky were actually ahead of their time, and had no equal then and none today.
By keeping an interest in AI, more eyes on the ball more players in the field. By staying true to the dream through the second AI winter. But if you are an average programmer with a reasonable grasp of algebra and algorithms play with some ideas, no one knows where the AI solutions are buried and you may get lucky!
The average person probably isn't smart enough to become an AI researcher, nor are they rich enough to fund AI research. I guess vote for politicians who support AI research and spread the word. Right now China is investing heavily in AI, which means the field will advance rapidly in the 2020s.
I think the mind is mostly unconscious. Try looking at a photo of someone you recognize and not think of their name. Or when you learn to drive a car or revise for a test, Are you just programming your unconsciousness?
I am mesmerized by the output of this two wonderful scientists, but i can't hide my discomfort by the insertions of personal convictions about consciousness, and the nature of life :/
😊 time belongs to compound compound belongs to my body 😊 I am a running portion of thoughts that cannot run on the time that is running through my body 😊 Therefore we just structure time to the time we exist
I want people to realize how mouth-breathingly reddit this video is: the guy that balks at the fact that people are upset at the idea that the brain is "just a machine" ... also thinks it's not even a machine and that you can copy it - what an absolute lunatic state of affairs people like this should shock the world and upload a human hair, toenail, skincell or kidney to the internet first, before they laughably talk about uploading a brain.
it staggers me that this guy both calls humans "machines" with delusions of spirit and self, and then questions the morality of murder based on the ability to copy paste his hypothetical machine allegory of a person. what if the "self" is the flow state that travels between these distinct brain operations? and if killing a copy is fine why are their moral constraints on murdering a single identical twin?
So the main takeaway for me is that Marvin Minsky believes that technology may surpass human intelligence within our lifetimes. And Minsky says that the thing we need to focus on in order to emulate human intelligence is architectures for high-level cognitive management that switch on and off different types of thinking and resolve conflicts between various processes in order to achieve goals. That approach described by Minsky sounds very similar to what IBM does with Watson and also the general approach of OpenCog. So I guess that people listened to his ideas and that is paying off.
I push people's buttons. I heard a sound I didn't like at hungry valley. It sounded like a bobcat or small cougar asking for help. That's what people say when they see mountain lions or bobcats right before they get killed and bobcats and cougars are both vocal enough to repeat it to get people to get out of their cars looking for a child asking for help then they kill them. I recognized it right away and I had the window by ear set up to only take one shot from a cougar. The way cougars speak it sounded closes enough to get to that one shot in a single leap. It could of been a person or the wind or even the cellphone off pitch and panned. But it sounded exactly like a young cougar asking for help. Bears and wolves can ask for help to but they sound different. I had a bear in my tent when I was 13. I got 10 feet away before I realized it was a bear. Luckily it was a 200 pound one that just got to eat my candy. I was 125 pounds and with me and the other guy I was with and our hands up and screaming to get out of there he left right before we did. He came back the next day why me and 6 other kids were eating dinner. He got within fifty feet before I spotted him. All 7 of us grabbed steak knives we had fitted into spears the night before and chased that motherfucker till we couldn't see him anymore and we slept with our spears all night for the next few days. We practiced and the seven of us together had it beat as long as it's mother didn't show up. Be had whistles cayenne powder and food to barter with. A fed bear is a happy bear unless it's a mother with a couple in trouble. At 13 it was rather a big deal but we all survived barely. The one kid had an asthma attack and panic attack at the same time and almost died
Minsky was no doubt a very intelligent person and had thought of a great many important things. But even the smartest minds cannot know everything. I refer to one peculiar statement he makes in the minute following 17:48. He talks about the need for AI in the context of the expected increase in human longevity. He suggests that AI will be necessary because people may live, say, to be 200 years old, and they would have to "have only one child person because the planet can't stand too many people..." The unstated assumption is that the resources of the planet are limited. This is in principle absolutely true since the plant is finite and therefore the amount of resources is finite. The mistake that Minsky makes lies in not recognizing that in practice (not in principle) the planet does not have any meaningful limit to the amount of resources it has.
At 5:50 Minsky says “some part of the brain notices that another part isn’t working so what does it do? It says, ‘oh I’ll turn some of those off and some of those on”. He implies a decision maker. How does that awareness arise that he is calling “I”? Somehow it would have to arise out of the neuronal tissue itself. That is a serious explanatory gap. He is not taking into account the implications of what he is saying. The “hard problem” in other words.
The term "Artificial Intelligence" is inherently problematic in that, there is no such thing as "artificial" intelligence, only intelligence. Trying to understand intelligence in terms other than universal will always present obstacles. Let me suggest the term "Actual Intelligence as an alternative" then you get to keep your AI. That will help admin costs. In any system, Intelligence is either present or absent and is universal in its form, function and application. Trying to understand intelligence as anything less than a universal constant is to anthropophasize and limit it to human experience. As a universal constant, Intelligence has fixed principles. This is a new, unexplored experience in human cognition and would yield much if treated as such.
Minsky's "empty box" in which a vacuous self-sense goes that is the ultimate overseer of brain processes is not vacuous as much as it is a radically empty plenum. Human longevity'll be increasing very rapidly, by changing a handful of genes. People can live to be 160 years old. Machines will be doing alot of the work people now do, but to be careful putting machines in charge of other machines and people before the bugs have been fixed.
I think the biggest problem with the baby learning programs is that hardware is currently incapable of mimicking our brain's neuroplasticity and growth. Once the hardware caps out, the computer stops learning.
This guys knows his stuff! Although I don't really agree with his arbiter mechanism theory as a separate neural entity as I see it as an inherent inhibition process of the neural networks, I think he's spot on on the rest.
right! and his reverse inhibition theory goes against the common admitted axiom:"What the fuck is this? How about the inference engine which leaks its oil?"... Fed up with jargon...
He has some interesting ideas but until we have a solution we wont know how far away he was from describing a working system, That's assuming that we find the solution he is describing and not something that's different but still works (or indeed if we ever find the solution.)
If I recall, I think the difference is more about timeline(s). Minsky expressed a degree of skepticism about a 2030 or 2040ish singularity. However when you get down to it, that is really the major point of disagreement, not IF but WHEN!
Should we pursue artificial intelligence is a less interesting question than is there any way to stop that pursuit. Personally I don't think there is any way to prevent the singularity if indeed it is a possibility.
i find this area fascinating. however, i fear that one day people will fall victim to companies that "upload" you to a computer and robot, and the real you -- body and soul -- die.
Genomic information is the bottom of the building blocks of the biological complexity. It does not reflect the organism's unique phenotypical ever-changing expressions and the evolution of learning of the organism.
Very valuable conversation indeed. However I cannot agree with them that we humans are only "machines", with no spirit/soul, etc ... For example there are thousands of videos even on youtube on various ocult, spiritist, prophetic and similar instances, that obviously show there is an extra reality behind mere physical and material one. I have even experienced some things myself. Almost no scientist like to talk about that, but it must be huge part, that we do not yet know or understand and cannot mesure with current methods. There lies the true revolution.
That is not what was said. The "soul" or "spirit" was mentioned in that video, and the label was ascribed to a general 'box' that people use for things they cannot explain about behavior, the architecture of the mind. Minsky said there was nothing profound about this, and I would agree.
Ydonkov that's why 500 years ago they also believed in spirits and burned on the stake anyone believing different. Any notion of soul is irrelevant. I count my atoms, they are 1 octillion. Souls are a fairytale and nowhere to be found.
What is his take on consciousness ? Does he think it is beyond our understanding or that what we perceive is just a model of reality and we can not access parts of it due to it only being a model so thus we may not be able to work out how the brain produces consciousness.. ?
@@kevinfairweather3661 I don't know what Chomsky thinks about consciousness, but his attitude towards AI or ASI is negative. He's in the opinion that were centuries away from even getting close to human level cognition in computers.
@@djacob7 Thanks. Yeah i would agree. We have systems that are great at completing distinct tasks but we are nowhere near building anything as flexible as a Human brain !
okay, i'm at just 5.27, and minsky tells that there are no such general principles of the brain as in physics. I think the brain is not the point. The brain should be analogous to the universe, and then the general principles would be analogous to principles of neural behaviour. We only should know how one neuron behaves, develops, evolves, and that principle with time would give rise to modular, complex neural structures. Would not it? We're far from that in computational speed ,however.
I always lose it at the backup discussion. While I can understand how other people will see the backup as the same person (as to whenever the backup took place), doesn't something die with the most recently "live" copy? If I use a back up copy while I'm still alive, from the outside there will be two identical people even the last "live" copy and the backup will both insist they are the real one, but obviously there are two entities. Obviously "I" (whatever that is, perhaps what we call our soul) cannot be active in both entities simultaneously.
a backup and original copy will be like two identical twins at birth. So at first it seems that you wouldn't want to kill either because the other exists. I am personally ambivalent. But two arguments that I think are useful in providing additional views towards reaching a conclusion are: 1-(due to minsky himself from another interview, paraphrased) we wouldn't have a problem replacing a person part by part until after a while the new entity shares no physical part with the original. In this scenario the same "self" continues to live uninterrupted even though it's not the same original physical entity. So is it reasonable to speed up this process by copying the person into an new physical entity and deleting the original in one swoop? I don't know. 2-(idea due to myself): when you sleep, in a way, you die and are resurrected again since your "self" is not active during sleep. Now if they copied you during sleep, deleted the original, and the copy wakes up from sleep, would it know that it's not the original?! And would the original know that it no longer exists?! if we knew (which we don't but to make a point) that sleep is a process where your "self" is turned off, then your hardware (brain) is vacuumed, and then the self is activated again (exactly the process of making a copy), then would we object to the copy-then-delete-original procedure?! I don't know.
@@smittymcjob2582 my point is identical twins at birth aren’t the same person. Neither is a backup. The idea that immortality can be achieved this way is nonsense. An android or whatever else with your exact personality is still not you. You still die and miss out on the infinite future
Having a back-up copy sounds like a cool idea. But honestly I don't understand why Kurzweil isn't disturbed at the idea of "you" living on without your consciousness being there, to me that isn't truly reaching immortality.
Xm Flash I believe that unless your mind is some how connected to your "backup" Its not really "you". But if you could say. Control 2 or more brains at once. Your consciousness resides in multiple brains instead of just one. So even if you kill one. You consciousness is still intact. I have a problem with the way they describe backups as well. By that logic. If i create a duplicate biological him right now. And place "him" right in front of himself. He wouldn't mind dying since that is still technically him? Horse crap.
+Rohan Mehta I take it he meant that if there is any strong AI it's probably in some other solar system, thing that I consider very plausible too. That or he just threw a quite obscure quote from some sci/fi novel he likes very much. ^_^
Minsky's ideas are fascinating and perhaps prescient. But they'll need to be articulated by a better representative. Like far too many intellectuals, he's smug and pretentious. And his speculations could very easily turn out to be wrong.
i dont agree: it is far much easier to replicate psychological models in a computer than to replicate the brain hardware in electronics. One of the most striking errors in the modern AI is exactly that: trying to replicate in SW the human brain ( A. neural networks) instead of taking psychological models into pure mathematic algorithms.
Have to take what he says with a grain of salt simply because he says human longevity will accelerate. He shows that he fails to take all things into account. Right now, the massive exposures to microwave radiation from wireless devices is causing major health disruptions. So is exposure to environmental toxins such as glyphosate. These guys may be smart but they are obviously oblivious or short sighted about many things too.
i had a similar thought. he talks as if everybody on this planet benefits equally from new discoveries. when reality is that even today we don't manage to give everybody access to food, shelter and health care while at the same time a certain percentage of humans wastes resources as if there was no limit.
I've been very much ok with the idea that I was a machine since I was let's say 14, I don't see how that's a problem to me. My shrink and kin may think otherwise since it of course affects my daily reasoning and therefore my behaviour but definitely it doesn't scare me to think I have no soul and that a copy of myself would be another myself plain and simple. I would love to see the day where I can copy myself an see how I evolve in a different spacial arrangement (since well we can't both occupy the same space) and different situations, it would be the most enriching experiment ever just talking to myself and see how our thinking ways part through the days and years... Just imagine the leaps in psychology we'd make if we could do that.... :D
What is this "myself" you are talking about? you are losing cells in your brain and skin every second. There is no stable state of "yourself" so how exactly do you copy "myself"?
Imagine that you create your copy exactly in the same position as you are but centrally symmetrically opposed to you in a room that is as well. This would be so that your brain inputs are the exact same for both. Due to this, your brains would work the exact same way and you'd probably do and say the same stuff at the same time, making your copy feel exactly like a mirror reflection (albeit a centrally symmetrical mirror). Which means that the image is nothing else but yourself in every sense or at least from your point of view. You know exactly what your clone is thinking because you are as well. He/she is yourself in all regards. Plain and simple. This would progressively stop being as synchronized through time due to quantum fluctuations (sorry if that's not the term), and you would see your mirror image eventually do something slightly different. You'd both be astonished in rather the same way and the asymmetry would only exponentially accentuate with time, but compared to other humans, that clone is just a "parallel universe" you, and no other has remotely the same experience, memories, intelligence, feelings as you do. Almost like you view yourself when comparing yourself to your yesteryear self. That's the myself I'm talking about.
The moment you create an exact copy of yourself, the particles in your copy wouldn't behave the same way as the original you due to Heisenberg's uncertainty principle. Your copy would probably disintegrate instantly since you have no control over particles' momentum.
Disintegrate instantly? I hope you can see how that's hard to believe. Especially since I've never seen a scientist disagree that teleportation was theoretically possible. Where can I look more into that idea of disintegration?
Are you familiar with Heisenberg's uncertainty principle? It states that the standard deviation of position of a particle times the standard deviation of its momentum is greater than some constant, which means when the position has less variance, momentum becomes more unpredictable. When you have infinite position precision, the probability distribution of momentum is stretched out to be a uniform distribution, i.e., equally probable with any value, hence unpredictable. (see en.wikipedia.org/wiki/Uncertainty_principle) I guess in layman's term, that means when you place a particle at a position you desire with high precision, you have no control over where the particle is heading or how fast it will be. The chance that all those trillions of particles move together to maintain a human body is essentially 0.
***** I definitely believe it's both a habit and for visualization, since I'd bet he's a guy who does a hell of a lot of trying to explain things most people have difficulty grasping.
***** could those hand movements simply artifacts from his specific brain attempting to retrieve information for his discussion? I doubt he is consciously moving them.
The increase in life expectancy is due to decrease in infant mortality, and the extrapolation to live more than 90 years as average is not supported by current data. But, well, are not all singularity talks just extrapolations without understanding?
That's a myth. Infant mortality has become a negligible factor to life expectancy in developed nations for several decades now, so the increase in the timeframe Minsky is talking about cannot be attributed to child mortality
BS. You need to understand transistors to understand how a computer works. Everything is reducible. You can understand the brain and consciousness from studying neurons, the connectome and chemical reactions within the brain.
