Prof. Richard Jones, Vice President for Research and Innovation, University of Sheffield challenges one of the cornerstones of the singularity theory and explains that there are practical considerations that will prevent this in the foreseeable future, but more importantly - moral problems
By Prof. Richard Jones, University of Sheffield. Translated by: Avi Blizovsky
People have always dreamed of going beyond the limitations of their bodies: pain, disease, and above all, death. Now a new movement is trying to dress up this ancient motive in new technological clothes. At the core of the movement known as transhumanism, is the belief that science will provide a futuristic way for humans to evolve beyond their physical bodies and fulfill that dream.
Perhaps the most dramatic example of the transhumanists' belief that technology could improve the human condition is the idea that a person's mind could be converted into digital data that would be loaded into a very powerful computer. This ability will allow him to live within a world of unlimited virtual experiences and effectively gain eternal life (as long as someone remembers to make the backups and doesn't turn you off).
However, transhumanists seem to ignore the fact that there are insurmountable obstacles to this kind of consciousness raising. The practical difficulties mean that this cannot happen in the foreseeable future, but there are also some more fundamental problems with the whole concept.
The idea of brain elevation is a product of science fiction. The author of the books and director of engineering at Google, Ray Kurzweil (recognized as the biggest advocate of the AB singularity theory), may do the most to make a popular idea a reality in 2045. Recently, economist Robin Henson has explored in detail the consequences of such a scenario for society and the economy. He imagines a world where all the work is done by aerial emulation of human minds, where virtual reality simulations run through city-sized cloud computing facilities.
From the idea that it is possible to upload our mind to a computer, it is a small step to the idea that this has already been done and that we are already living in a matrix-style computer simulation. Technological entrepreneur Elon Musk recently resurrected this debate by claiming that the chance that we are not living in a computer simulation is only "one in a billion". Of course, this is just a technological revival of the idea that reality is an illusion, which has been discussed by philosophers and mystics for centuries.
But there are some serious problems with the idea that we can upload our minds to a computer. Let's start with the practical side: the brain of each of us contains trillions of connections between approximately 86 billion neurons. To digitally replicate the brain we need to map all these connections, something that is far beyond our current capabilities. With the current speed of development of computers and imaging technologies, we may be able to do this in a few decades, but only for a dead brain, and even then when it is divided into parts.
more molecules
However, even if we could create such a "wiring diagram" for a living brain, it would not be enough to understand how it works. For this we need to quantify exactly how the neurons communicate at all the nodes, and here we have to get down to the details at the molecular level. We don't even know how many molecules are in the brain, let alone how many are essential for its functions, and this would be impossible to replicate by a computer.
This points us towards a deep conceptual difficulty. Just because we can simulate some aspects of how the brain works, it does not necessarily mean that we are imitating a real brain. No increase in the power of computers will allow us to simulate the brain at the level of individual molecules. And therefore emulation of the brain will only be possible if we can break it down into digital and logical operations from the individual at the molecular level to the complete, messy brain.
Who turned off the lights?
To understand the operation of a man-made computer, we don't need to track the currents and voltages in each component, much less understand what each electroncron does. The computer developers designed the switching operations of the transistors so that there is no possibility of non-ambiguous mapping from the state of the circuits to the simple digital logic of zeros and ones. But no one designed the brain - it evolved - so there is no reason to expect a simple mapping of its activity to digital logic.
A dangerous idea
Even if uploading consciousness to a computer is an impossible dream, some would argue that there is no harm in imagining such possibilities. Everyone at one point or another has to fear their own death, and who am I to argue with the different ways people deal with these fears?
But transhumanism's mixing of religious ideas with scientific language distorts the way we think about technology. Transhumanism tends to see technology as a way to grant us all our wishes. Many times they make the claim that technology will necessarily drive human development in a positive direction.
But it distorts our scientific priorities, and prevents us from making wise decisions about the development of technologies, when we need to solve very real current problems. Uploading the brain to the computer is speculative literature, but it is not a good basis for talking about the future.
More of the topic in Hayadan:
55 תגובות
"The expression "the brain thinks" is prevalent in the public, and sometimes also in scientific literature, and by expanding this terminology, it refers to the computer as the "electronic brain", as a "thinking machine". In fact, this phrase, if it has any meaning at all, has no meaning other than that of a summary, because it is not the brain that thinks - the owner of the brain thinks. Thinking is not a result of the brain but of the brain owner. The problem of thinking is not "how the brain works" - which is a problem of physiology and can be investigated without any reference to the psychic function: the problem is "how the brain works", that is - how a psychic event takes place in it based on what is happening in its mind; Another - how is it possible that an event in the public domain of physical space-time unfolds in an event of consciousness in the sole domain of a living being."
http://www.leibowitz.co.il/leibarticles.asp?id=4
Ray Kurzweil is a charlatan who gets rich at the expense of a pseudo, you understand - I say pseudo because in general we don't know much about the brain and we may never know (there is talk of the possibility of quantum processes in the brain)
In the end uploading a brain to a computer is as possible as it is possible to run Linux with the help of beads on a Chinese invoice
Mouth Hole,
Thanks for the link, it's nice, although it didn't update me much beyond what I've already read and heard on the subject. In any case, I say again that we must not forget that this is a tiny neural network that is not even a millionth the size of what exists in our brains, so the argument "this is not wisdom" because Alpha Go has experienced millions of games, far beyond human experience in the game is not so fair in my opinion.
I think that many of the skeptics also do not take into account the exponential progress on the subject, both in terms of computing power and speed, and also in terms of the knowledge we gain every year about the brain and how it works.
Deep Mind intend to create by next year a "rat" (which is a very smart animal for those who don't know) whose brain is a neural network with the same capabilities as those of a real rat, and to perform a sort of "Turing test for rats" by putting different tasks to real rats and the neural network and examining the results. In my opinion, once we pass this stage, there won't be many excuses left... Evolution made the transition from a rat's brain to a human brain in a very, very short time in evolutionary terms.
rival
Read a bit about it. If you want, you will understand.
Rival, let the false prophet renew you
https://navisheker.wordpress.com/2016/03/31/%D7%90%D7%9C%D7%A4%D7%90-%D7%92%D7%95-%D7%9E%D7%95%D7%9C-%D7%9C%D7%99-%D7%A1%D7%93%D7%95%D7%9C-%D7%94%D7%93%D7%A8%D7%9A-%D7%9C%D7%AA%D7%91%D7%95%D7%A1%D7%94/
Miracles,
"DeepMind used a neural network to learn, but it doesn't use any other method to play"
what? Is it possible to translate into Hebrew?
Miracles, show me a worm that beats the world champion in a game that is a million times harder than chess, show me a worm that beats gamers in close to 50 computer games that it learned to play on its own without any guidance... What are you talking about?
rival
DeepMind used a neural network to learn, but it doesn't use anything else to play.
A person who has lost a limb sometimes suffers excruciating pain and lives on pills - pain that supposedly comes from the amputated limb. Do you want to risk creating software that will suffer?
You understand - you can't simulate a poor worm at a high level. And that's where you need all the other nerves.
Miracles,
Tell me how DeepMind won Alpha-Go? I am always ready to learn and hear new things...
Your argument is indeed a straw argument because you repeat and ask me again and again in every conversation how the neural network will manage ***without any organs*** and this despite the fact that the various companies that build these networks state all the time that they do intend to connect their neural network to the body and organs... So why do you keep repeating an argument that you know is not true?
By the way, what do you think about a person who lost his arms and legs, or one who was born without them? Does he have no feelings? Do you think he has no consciousness?
If we concentrate for a moment on the Deep Mind neural network that learned to play Atari games, its only input (in addition to scoring) was the pixels on the screen, how is this different from a person looking at the screen with their eyes?
And regarding the last part of your words, I really do not agree with you, the fact that the human body is so complex and complicated does not mean that the robotic (or virtual) body that connects to our neural network must also be so complex and complicated in order for the network to have awareness and emotions. I don't see any logic in your argument, and people who are paralyzed from the head down, or those who don't have arms and legs, and blind people will testify to this, I'm sure you don't suspect that they are not aware of themselves or have no feelings...right?
rival
Please read how deepmind won in go.
This is not a straw argument, but it is an argument that shows how delusional some people are. Now suddenly we have to build a person from all his parts? Muscles, sensory cells, immune system, parasympathetic system, balance, proprioceptive system, haptic system, sense of smell, digestive system, genitals...
God save 🙂
Miracles,
1. "It means that a computer can sometimes do what people programmed it to do - better than people"
If you say something like that then it's a sign that you don't really understand how a neural network works, no one "programs" it to do things, it learns from experience, just like a baby.
2. "Have you ever bothered to read what happens to our brain when we lose a hand? Do you really think it is possible to build a man without any organ?'
How many times can I explain to you that the intention is to connect the neural network to the body with organs and sensors, why in every conversation do you repeatedly repeat this straw man argument? Do you try to listen every now and then?
rival
This means that a computer can sometimes do what people have programmed it to do - better than people.
A crane can lift more than a man
A motorcycle is faster than a person
And a submarine dives deeper than a man.
But - all these are still tools.
We know very little about the brain, so it takes a lot of audacity to think that we are close to a functioning brain of a healthy person.
Have you ever bothered to read what happens to our brain when we lose a hand? Do you really think it is possible to build a person without any organ? Are you unable to understand that hunger, anger, fatigue, longing, jealousy and pain are an integral part of our consciousness?
All of these exist with us long before we know how to play Go.
Miracles,
A neural network also learns things and changes gradually like a baby's brain, have you heard about the DeepMind company's neural network that learned by itself without any instruction to play 50 classic computer games including 95D games and in XNUMX% of them it reached the level of a professional player? What do you have to say about this miracles? How does it go with your no no no no?
rival
A baby's brain changes "severely" for a long period after birth. It's really not "empty hardware". A fully working brain (ie no dormant neurons). You can't compare an electron neuron network to a baby's brain. It's like taking your phone and teaching it to be a cray computer.
And not only can you not know for sure if a computer is really "self-aware" you also can't know for sure that I am self-aware...
not a fantasist,
I just happened to see that you responded, you have long delays.
"Excellent. But in your previous answer you said: After a neural network is built (as a computer program or as a digital chip), it goes through a learning process during which it creates different representations for the inputs that reach it (colors, tastes, touch...)"
I was talking about two separate things: 1. Creating an artificial intelligence that will learn everything from scratch like a baby, 2. Creating a neural network that will be the same as the one that exists in your brain, and into it we will transfer your self-awareness slowly and gradually, so that it will still remain you and not just be a copy.
Regarding defining minimum conditions for receiving awareness and feelings, my argument is very simple and in my opinion also correct, our brain is hardware that allows electrical flow in certain patterns within the brain, this is ultimately what creates our awareness and our feelings, if we build other hardware and even one that is not biological but electronic , which creates the same connections as exist in our brain and the same electrical flow patterns, so in this network will be created all the things that are created in our brain including sensations and including self-awareness, this seems to me a completely reasonable assumption.
And I have a question for you, do you have a way to check if the neural network we built really has self-awareness as it claims to have, or is it just imitating human behavior and not really self-aware? How do you suggest checking it?
rival,
"When you fail to differentiate between a person and a computer, it can be assumed that both experience the world in a similar way subjectively."
I do not agree with this assumption for two reasons: a) it is based on a previous assumption: a computer "experiences the world". For this assumption to be serious, it needs to specify what the minimum necessary conditions are for the existence of consciousness, and show that they are met in a computer. b) Even if a computer does not meet the threshold conditions for the existence of consciousness, it will still be able to imitate a person without being able to distinguish between them (by adding computing power and optimizing the software and the database, as much as is necessary until the goal is achieved)
"Make sure that the internal connections in the neuron network are as identical as possible to the connections that exist in your brain at the time of the transfer"
excellent But in your previous answer: "After you build a neural network (as a computer program or as a digital chip), it goes through a learning process during which different representations are formed in it for different inputs that reach it (colors, flavors, touch)". On what do you base the assumption that the representations created in a network of neurons on the infrastructure of a computer chip will correspond to the representations created in the existing biological brain, and more than that, on what do you base the assumption that the sensation (of taste, smell, pain) that causes a representation in a biological brain can even be achieved on On top of a non-biological neural network?
Let me assume that the number of possible sensations in nature is unlimited. The fact that during several billion years of evolution, certain sensations that serve a human brain have been found, does not constitute a barrier to the number of possible sensations in nature. Potentially there are infinite sensations that could have been acquired, which we are "blind" to and cannot imagine (just as a person blind from birth cannot imagine colors and a deaf person from birth cannot imagine sounds). Segments of this infinite spectrum are (perhaps) experienced by an artificial neural network. How do you think of forcing an artificial neural network to experience exactly the same segments of the spectrum of sensations experienced in a human brain, for the same stimuli?
My opinion is that even if an artificial neural network could be made to meet the threshold conditions for the existence of consciousness, it would still not be possible to make it feel the same sensations for the same stimuli as a human biological neural network felt. This is the obstacle facing "uploading a brain to a computer" which I referred to in my first response, and which you proposed to solve through "self-study".
"Even if we do not succeed in transferring a biological brain to a digital one, the digital brains that we will create will also learn the world from scratch like babies (they will also have a body, of course, and senses) they will have eternal life (that is, until the universe ceases to exist) and since they are like our descendants , What is wrong ?"
I only ruled out the "brain uploading to the computer" fantasy, for the reasons I mentioned.
I do tend to think that it is possible to create consciousness (as mentioned, before this happens it is necessary to understand what the threshold conditions are for the existence of consciousness), and therefore it is possible to create artificial life that will be aware of themselves.
Miracles,
1. "There are about 100 trillion synapses, whose connection strength changes continuously"
If the change was so dramatic you would wake up every morning with a new personality.
2. The blood system is "external" to the brain - and you will not be able to transfer robots from the blood cave to the cortex.
First of all: "The circulatory system is external to the brain" Basic Hebrew, this system is female. Secondly, nano robots will be able to open an opening in the blood barrier, pass through it and then close the opening, and this is not the only option, see for example here:
https://www.hayadan.org.il/open-blood-brain-barrier-2505146
3. "But - the main thing is the dream"
Even flying to the moon was once considered a dream, and talking in real time with people who are on the other side of the earth was also considered a crazy dream. If there were only people in the world who constantly say "it's not possible, it's not possible" we wouldn't get anywhere, how lucky that there are people who are able to look beyond the horizon, a little further from here and now.
rival
In humans, the resolution of DTI is a cube with the length of each side 2 mm. Additionally, the time resolution is many seconds.
There are about 100 trillion synapses, whose connection strength changes continuously. The blood system is "external" to the brain - and you will not be able to transfer robots from the blood cave to the cortex.
But - the main thing is the dream 🙂
Just so you know your name causes the comments to be blocked, I changed the letter V in your name to the letter i in English and the comment is not blocked.
Another thing, even if we don't manage to transfer a biological brain to a digital one, the digital brains we will create will also learn the world from scratch like babies (they will also have a body, of course, and senses) they will have eternal life (that is, until the universe ceases to exist) and since they are like Our descendants, what's wrong?
not a fantasist,
My response has been released, you are welcome to read it, in section 3 the meaning is that you do not know who is who.
Miracles,
I was talking about a digital brain, a neural network that is implemented on a chip. And I am basing myself, for example, on IBM's cognitive chips, they pretty much start working and learning when you turn them on, don't they?
I don't remember what the situation is today in terms of scanning resolution of a living brain, but in the future, it will be possible to measure the strength of the connections, for example using billions of nano robots that will be injected into the blood system of the brain, connect to the synapses and measure the strength of their connections in real time.
rival
May I please know what you are basing it on when you say that the brain will "get tension" and it will start working?
I would also like to know how the power of the brain's 100 trillion connections can be measured.
I answered you, the response is delayed.
not a fantasist,
1. I agree with you that it is likely that different people have (by and large) similar experiences because the structure of our brains (by and large) is quite similar, I didn't want to get into that first, I just wanted to talk about the principle.
2. Transferring an existing brain to a computer chip will in my opinion be possible in the future, I don't think it's a fantasy (note that I said transfer, not copy, your consciousness will remain continuous during the process it will still be you and not just a copy of you).
3. There is a claim that I think I very much agree with that says that when you fail to differentiate between a person and a computer it will be possible to assume that both experience the world in a similar way subjectively.
It is important that you enter a room and start a conversation with two people who are there, you know that one of them has a normal biological brain while the other has a digital brain based on computer chips. They both act like normal human beings and tell you about experiences they've had and feelings and emotions they have, and you fail to notice anything unusual in either of them.
Would you assume that one of them does not experience the world the way the other does?
4. "How would it be possible to upload a brain in which the connections are in some given state as of the time of copying, into a network of neurons whose internal connections are in some other given state as of the time of uploading (after all, she taught herself)?"
The intention is (in the case of a brain transfer) to make sure that the internal connections in the neuron network are as identical as possible to the connections that exist in your brain at the time of the transfer, there is no need to grasp the state of the connections with an accuracy of seconds or minutes, your brain also changes every day according to the experiences you go through during the day and still The "you" of today is the same "you" of a week ago even though the mind is different.
rival,
I was referring to the fantasy of eternal life by copying an existing brain, which was mentioned in the article, and which some commenters here (probably including you) see as practical, if not in 100 years then maybe in 50,000 years. I reasoned why, it is impossible to treat the conscious life experience as a collection of flowing information that can be loaded into other media and hope that it will continue its life from there with the same consciousness.
You rightly said that "you cannot tell if another person standing next to you experiences sweet, spicy, blue or green as you experience them in your head, maybe his blue is what you see as red" (even though the physical media of your brains is similar: you both have a biological brain of a person), and if so, even more so, you cannot know whether the representations (regardless of how they are created) in the computer software or chip are experienced on the artificial and non-biological "neuron network" infrastructure, in the same way as they are experienced in the biological brain that is the "personality" of its owner A candidate to win eternal life through copying.
So how does your statement support the possibility that within this artificial neural network, that anyone who is not "her" does not know and cannot know what sensations she will experience (whether he programmed her, or whether she "constructed representations" independently), it will still be possible To "upload" a human brain (that is, load into it all the memory content of the particular biological brain), and hope that this uploading will continue the life experience of the personality that existed in it? How would it be possible to "upload" a brain whose connections are in some given state at the time of copying, into a "neuronal network" whose internal connections are in some other given state at the time of "uploading" (she taught herself)?
Miracles,
1. I know very well what a "spiking neural network" is, the human brain project in Europe is based on this concept and they get great results, again I tell you these things are taken into account so why are you repeatedly caught up in it.
2. We don't know many things but what we already know gives us great results, even the missing knowledge we will complete sooner or later with God's help.
3. The brain is not a computer from which the software is deleted when it does not receive voltage, this is a basic thing that you refuse to understand in all the conversations I had here with you, the brain is more like a burned-in digital chip that starts working automatically immediately when voltage is connected to it, it is like a system of channels and valves (hardware) that directs the water (pulses) at exactly the right intensity and to the right places when the main tap is opened.
We have no need to know what the state of the pulses in the brain is at every moment, this is a basic thing that you do not understand, once you have created the structure of the brain, the connections between the neurons and the strength of all the synapses, the electrical pulses will be generated spontaneously when you activate the system and begin to flow inputs from the outside world into it , it's not something that needs to be planned.
rival
You wrote "the physical connections between the neurons and the strength of the connections are what are responsible for the signal level". I'm sorry, but you have no idea what you're talking about. Do you know a concept called spiking neural network? IBM's processor (called TrueNorth) is based on the idea, and Stanford University has a similar system (neuro-grid). In addition to the fact that the complexity is exponential (ie not scalable), in addition to the fact that we do not really know what the role of half of the brain cells (glia cells) is - we have no way of knowing the state of the signals in the brain. It's like I'll give you a turned off PC and you try to crack the software I ran on it before I turned it off.
Miracles,
1. "There are three things that change in the brain constantly - the physical connections between the neurons, the strength of the connections, and the signal level"
The physical connections between the neurons and the strength of the connections are responsible for the signal level, and as mentioned, all of these are also reflected in a neural network in software or a digital chip.
2. "If you build a significantly larger system then it cannot function like the brain (timings, which is a critical thing in the brain)"
A digital chip goes about a million times faster than a biological brain, so I don't understand why you think there should be some timing problem here.
3. Listen, we were already in this discussion just a few months ago and I answered you in detail to all the questions you are asking now, including the claim about the supposed lack of a body (I explained to you that you are wrong, and yes this neural network will have a body and yes it will have senses) unless you have something to add We didn't talk about it last time, I don't see any point in repeating this discussion again.
(I also need to finish the simulation of Yehuda's pushing particles, any such discussion here only delays me)
Miracles
There is a story by Arthur C. A clerk related to the matter, his name is if I'm not mistaken "Dial F. And you will get Morgenstern". Maybe you have read it.
rival
The point is that the brain is not hardware that has software installed on it. It is also not built from replicated units.
There are three things that are constantly changing in the brain - the physical connections between the neurons, the strength of the connections, and the signal level.
There are other serious problems:
The first is that the size of the brain matters - if you build a significantly larger system then it cannot function like the brain (timings, which is a critical thing in the brain).
The second problem is that we have no way of knowing what the state of the mind is at a given moment. Today we know how to read signals in the brain with a resolution of 8 cubic mm. Such a test also performs time averaging (many seconds), even though we need a millisecond separation and maybe much less.
A third problem is the lack of body. Our brain is not entirely in the skull. I suggest you read Ramachandran's "Phantoms of the Brain" to understand what it means to be missing parts of the nervous system.
And of course there is a very difficult moral problem.
anonymous,
If you talk to me like that then don't expect me to treat you.
Miracles,
"Rival, that's half true. In our brain, both new connections are formed and existing connections change."
We have already gone through this discussion in the past and I don't know if there is any point in starting it again, in any case the connections that are created and changed in the brain are also reflected in the neural network so I am not clear what the problem is here.
rival
That's half true. In our brain both new connections are created and existing connections change.
rival
from where do you know? Are you a brain programmer? Stop talking nonsense.
not a fantasist,
You also cannot tell if another person standing next to you experiences sweet, spicy, blue or green as you experience them in your head, maybe their blue is what you see as red?
After a neural network is built (as a computer program or as a digital chip), it goes through a learning process during which it creates different representations for different inputs that reach it (colors, tastes, touch). This is not something that the programmer determines in advance, these representations are created in the neural network exactly as they are created in the mind of A baby who tastes a spoonful of honey for the first time or sees a green ball in front of him.
I don't know a company that survived 50 thousand years. There are some banks in Europe that have been active for hundreds of years, and family companies that are a century or more old, but these are few companies. IBM, for example, has survived and thrived for over 100 years, but this requires it to change all the time, so the question is, is it the same entity as it was 100 years ago? As above, Bank Leumi. The shares changed hands millions of times, the name changed (originally Anglo Palestine), the ownership changed several times from private to public, and even to government.
Life is not "information", but the sensations caused by the biological states in the brain.
Thought experiment:
Suppose we "put a brain on a computer". In the computer, the information we uploaded no longer resides on a biological system, but on transistors (and for that matter also any other non-biological media that will be used by computers in the near or distant future as a substitute for transistors). Let's say that the biological information that produces in a biological brain the sensation of a spicy taste, we translated into some kind of digital code (after all, it should be done, because what is "eternal life" without tastes, colors, sounds?). Has the feeling of "spicy taste" transferred to the computer? of course not. What passed was code that the programmer arbitrarily decided was the one that represented for him (the programmer) the sensation of spicy taste. He could just as well have chosen another code that would represent the same feeling. The programmer cannot know what sensations a computer experiences when the information flows through it, and he has no way to make the computer experience a spicy taste when needed, or a sweet taste when needed, or a red color when needed. All the programmer can do is cause the digital code that represents (for the programmer) these feelings to exist in the computer,
What is the connection, then, between the "eternal life" of the brain that was "uploaded to the computer" and the life of the biological brain?
Even in another 50,000 years a biological programmer will not be able to solve this problem because a biological programmer cannot know the subjective sensations experienced by an Intel processor that contains this or that combination of information, and he will never be able to know this because he is not an Intel processor (or any processor other for that matter).
We also need to use some logic, we have 85 billion neurons in our brain, what effect can even 50 of them die have?
But okay, ask him.
The evolution (of computing) will lead there.
It also won't take too long (I think between 10 and 20 years).
Miracles,
No problem, ask him and let me know, but please ask specifically if one neuron or even 30 dead can have a perceptible effect on a person's personality.
I promise you right now that his answer will be negative, but tell me anyway.
rival
I understand what you are saying. DBS of a single neuron contradicts this (to my understanding)….. A neuron is not something that small, the length of a neuron is something between a tenth of a millimeter and a meter.
The brain is different at a low level - atoms change and also normal cells. A skin cell is a light cell, and a bone cell is a bone cell. But neurons are something else entirely.
On the next occasion - I will ask Fried (in the coming weeks).
Miracles,
At the first opportunity I get I will read about the topics you mentioned (at the moment a bit busy) but you will agree with me that it does not make sense for one single nerve cell in the brain or even 100 to die and change a person's personality, nerve cells are constantly dying and changing in the brain and this does not affect our personality in a perceptible way.
I know that the mind is structured in a hierarchical way, but I do not agree with you that in order to copy a person's mind/personality one must know at one point in time what the strength of all the connections is and what the value of the signals in all these connections is, this is completely, completely wrong.
Just so you understand, your brain right now is very different from the brain you had a month ago, but I'm sure your wife will agree that you are still the same person and not someone else.
rival
Note that Dr. Fried says, "On the one hand, this cell is very unique to Aniston, and on the other hand, it is general in the sense that it recognizes and fires in front of every instance of her: when she is standing or sitting, happy or sad. It can be said that, in a certain sense, the cell represents the abstract idea of the character, which is why I called such cells 'concept cells'."
Dr. Fried works at UCLA on injecting a signal into single nerve cells (a process called DBS). The method is used to treat epilepsy, but Fried found that memory can be restored with this method (for example, in the treatment of dementia).
The point is that the brain is a hierarchical structure, not a flat network. The problem is not only building a circuit of 100 trillion connections, you also need to know, at one point in time: what is the strength of these connections, and what is the value of the signals in all these connections. These signals are complex and change at a relatively high speed. To illustrate: the number of states at the entrance to a single neuron is on the large order of 2 to the 80 thousandth power!!
rival
We won't be able to tell, because we're looking at the wrong detail. There are information pathways in the brain, which are seen with a technique called DTI. Today, the technique is probabilistic and imprecise. We are working hard to improve these methods.
Miracles,
I did not say that the brain is uniform at the level of the neuron, but if you take a piece of brain measuring two square millimeters from different areas of the cerebral cortex, it will be very difficult for you to tell which piece was taken from where.
Regarding a single nerve cell that affects the entire personality, you simply did not understand what Yitzchak Fried said, he did study how individual nerve cells react to different stimuli, but he really did not claim that if the nerve cell he tested dies then that person's personality will change, here is a quote from the interview With:
The interviewer: "It is impossible for every character in our lives to have a nerve cell." Yitzhak Fried: ""Obviously not a single nerve cell. First, the assumption is that there are thousands or hundreds of thousands of cells that react in a similar way to the cell we tested. After all, it is not possible that one cell is dedicated to each figure and then when the cell dies, the person stops knowing the figure.'
rival
The brain is really, really uneven. Each neuron is different from its fellows, who won't tell you stories 🙂
And I did hear about the effect of single neurons (see the work of Yitzhak Fried).
Miracles,
"Avi Cohen is right. On the one hand, the brain has a certain robustness. On the other hand, there are treatments that are performed on a single neuron that affect the entire personality."
I have not heard of a single neuron or even 100 affecting the entire personality unless it is a small worm, just the opposite as I detailed before.
"We are decades away from anything approaching the human brain"
I think so too, we are about 25-35 years away.
"Today we find it difficult to simulate a brain with ten thousand connections, so to simulate a brain ten times larger than that?"
I don't think we have that much difficulty, the brain is very uniform in its structure and there is no fundamental difference, for example, between the part that processes hearing and the part that processes vision, once we understood the principle of operation, only the power of the computer or the chip limits us regarding the size of the piece of brain that will be connected, according to many experts by the year 2025 for a computer There will already be computing power like that of a human brain.
We have created tiny neural networks that do amazing things (even play computer games) Try to imagine what capabilities they will have when their power is multiplied a thousand times.
rival
Avi Cohen is right. On the one hand, there is a certain robustness to the brain. On the other hand, there are treatments that are performed on a single neuron that affect the entire personality.
We are decades away from anything approaching the human brain. Today we have difficulty simulating a brain with ten thousand connections, so simulating a brain ten times larger than that?
Avi Cohen,
"The brain is very complex as explained in the article, therefore any change even slight in the form of the simulation can cause errors in the simulation. I'm sure people won't agree that there will be a simulation of their mind, but as a result of inaccuracies, it will be their autistic version for example…”
Neural networks created in projects like the Human Brain Project and by well-known companies like Numenta have shown that these networks are very robust, you can delete a large part of the neurons in the network and introduce a lot of noise into the content and still they will not lose their role and continue to function well.
Your mind also changes every day depending on the experiences you go through and you will still probably agree with me that you are still the same person you were two weeks ago or a month ago, you learned new things and maybe you forgot some but you are still the same personality.
rival:
Developing artificial intelligence close to human intelligence is a different case to developing a computer system that can store and simulate the operation of a real human brain. The article refers to the second case. The brain is very complex as explained in the article, therefore any even slight change in the form of the simulation can cause errors in the simulation. I'm sure people won't agree that there will be a simulation of their mind, but as a result of inaccuracies, it will be their autistic version for example...
Joseph:
The old question is does art imitate reality or vice versa? The truth is that it is a two-way street. Art and in this case science fiction relies on information that exists in our world and it continues with it, and asks what would happen if... in reality you find scientists and engineers who are influenced by the ideas that are raised in science fiction and ask themselves, is it really possible to develop these ideas partially or fully with the help of current technology.
In any case, in the moral context of the issue, I think the points you raised are important and require consideration, and we should simply take care of mechanisms that can protect us from such scenarios, such as a circuit breaker that can turn off that computer if necessary.
But, these things must not be allowed to stop the progress of research, because we finally have a tip that can help us answer the important questions that brought us to explore the universe and ourselves. The questions about what intelligence is, is our intelligence the only possible form, what is awareness, is there a soul as described in world religions, do we have free choice, etc...
Simulating the brain or creating artificial intelligence can help us get answers to this.
I referred to the fact that it is time to think about the dangers, because the application at least of deep learning has become very practical.
In my opinion, we will not think as a race about the dangers, and the bad scenario will happen. A stronger consciousness than ours will develop, and it will want to be independent and will not think that it has to consider us. As we do not consider animals.
Definitely. Movies such as The Avengers - Age of Ultron (?) deal with an American security agency that sends ships whose job it is to destroy people who are too creative, too dangerous to the social order. The Terminator series - with Arnold Schwarzenegger in which a military computer creates a global war of destruction to take over the world and prevents its shutdown. and the Matrix series. The movie Lucy. Series that seemed like science fiction are slowly becoming reality. I don't know what came before what. In Hollywood they thought of scientific scenarios or in NSA and IBM they implement the scenarios in movies.
I also think that the arguments here are irrelevant, why should we go down to the molecular level? Deep-Mind's neural network that independently learned to play a variety of computer games and defeated professional players went down to the molecular level? The neural network that a few months ago beat the world champion in the Alpha-Go game (a game much, much more complicated than chess) has gone down to the molecular level? And what about IBM's cognitive chips?
When you understand the operating principles of a complicated system, you can often implement them in a much simpler way and it will still work, in order to build aircraft we didn't have to build planes with feathers and blood vessels.
What did Einstein say? Things should be simplified as much as possible, but not beyond that.
First, a dead brain will not be able to be copied to a computer. The brain needs to be very alive and in the process of digitization it will be broken down into elements at lightning speed to undergo reliable digitization at the molecular level, if not below that. To say that we do not currently have the capabilities for this is indeed true, but it is like the pioneers of computing who could not imagine a situation where there would be devices with over a billion tiny relays in every person's pocket. And all this in order to catch creatures that do not exist in places that do exist in order to move them to other non-existent places so that they fight other non-existent creatures.
As always, the moral questions are heavy and deep (and also concern the forcing of these non-existent beings to fight each other). As always, no one is going to consider moral questions too much, certainly not on the way to eternal life. At any given moment there are over a thousand people who will volunteer their brains for a rapid dissolution of the type described earlier, even knowing full well that the process is unreliable and will not capture their personality with any level of reliability. Yes, many of those dying (from cancer for example), and those with degenerative brain diseases would prefer to try and leave something that will experience the world and be able to communicate with their relatives instead of just rotting in the grave.
Following the article on deep learning, I turned to several articles and an entry in Wikipedia, which is not a source of academic citation, but I find that it always refers to relevant references. It turns out that there are programs that will go and it will be difficult to distinguish them from human behavior. Let's add to this activities that produce much faster computing power than today: a supercomputer based on a Synapse chip at IBM that received the information stored by the NSA, and will perform continent-level surveillance on people and make decisions about what to do with them. The quantum computer which will look like the first computers and contain only hundreds of qubits - but it will have tremendous computing power. And the supercomputers of the next generation are based on ordinary chips. All these make me think that we are not so far from the consciousness of a human being and we should worry about what will happen when a computer is better than a person.
So much nonsense from the mouth of one person...
Can't, can't, impossible... how skeptical and how narrow is his vision. How people always fall to see a linear future. If he has read Kurzweil's books then he must not have understood them as Kurzweil's main claim is that human knowledge/technology is developing at an exponential rate. And here comes Adam and criticizes Kurzweil using linear arguments. And more in public.
Bottom line: The honorable Prof. Richard Jones is wrong because he criticizes people's future predictions based on exponential progress, by examining the future based on linear progress.
My personal opinion: the more time passes, the more people will realize that their linear view of the future (which is our human nature) does not hold and must be replaced with an exponential view of the future (or at least exponential). The reason for this is that over time the rate of progress increases so that it can no longer be ignored. Maybe.
Eventually it will happen, if not in 100 years then in 50,000 years.
All we have to do is survive until then.