Comment on this entry
Why Worry About This Sci-Fi Stuff Now?
The term “mindclone” evokes a wide range of sci-fi images from the “Cylons” of Battlestar Galactica to the “Mr. Smiths” of The Matrix. While it is indisputable that we are creating large mindfiles, as described in Question 1, and surely there are geeks working hard on mindware, as reviewed in Question 2, how close could we be to an actual mindclone when computers can’t converse on their own much better than a two-year old kid?
...
Complete entry
COMMENTS
Posted by Hervé Musseau on 05/22 at 08:46 AM
It's interesting to see someone use Kurzweilian projections, as even in the transhumanist community they are often deemed as being on the optimistic side (without being completely rejected either).
How do you compare mindfiles/mindclones to RK's reverse engineering of the brain and mind uploading? Do you see it as being a step prior, or as being the more realistic achievement?
Posted by Armand on 05/22 at 02:11 PM
One thing you don't consider is that your estimation of Human brain power is wrong. The brain is not a digital computer. Processing occurs not only on the level of the individual neuron, but also as the brain acting as a whole and on every conceivable level in between the two. There's a lot of processing done on the sub-neuron level aswell, involving neurotransmitters, hormones, and other molecules. Even if you accept Kurzeil's claim that 1000 computations a second would be enough to simulate a neuron (which I don't) this doesn't take into account the massive processing power that occurs both above and below the level of individual neurons.
Knowledge may be unlimited, but hardware ain't. Physical limitations will eventually slow down and stop the increasing power of computers. We are also nowhere near craking the neural code. We no doubt use mutliple coding techniques, and neural codes are not just unique to every individual but change over the course of a person's life. If we can't translate someone's neural code into machine code, how can we write the software for a mind clone? For these reasons (underestimation of brain power, inevitable failure of moore's law, and the inability to decipher the Human neural code and thus create mind clone software), I do not believe that mind clones are in the foreseeable future.
Posted by Athena Andreadis on 05/22 at 04:24 PM
The foreseeable future or any future, if you define them as backup copies of yourself. If they can be successfully created (an enormous if, for reasons that Armand pointed out), they will be distinct entities.
Posted by Giulio Prisco on 05/23 at 01:24 AM
They will be distinct entities in the sense that I am a distinct entity from that toddler who played with cubes.
I think after uploading (like after every significant change), I am still me if and only if both the previous me and the future me are willing to accept the future me as a valid continuation of the
previous me.
Posted by Athena Andreadis on 05/23 at 05:22 PM
After the previous you accepts that you are the future you, will you kill the previous you as redundant? You're not a distinct entity from the toddler you were. You should know better than that, it's Neurobio 101. Your brain neurons are almost entirely the same. If you get dementia or a stroke, you'll become a distinct entity as well -- and you (maybe) and those around you (certainly) will know you are no longer you.
Posted by Giulio Prisco on 05/24 at 08:20 AM
I will not kill the previous me as redundant, for many reasons, the principal being that I would not kill anyone as redundant. After the copy process and as soon as we begin to diverge, of course we will be two different persons in many senses, but also two equally valid continuations of the previous single person.
The toddler. Most brain neurons are the same, so what? The information encoded in the two sets of neurons is very different. I prefer to identify with the current information rather than with the hardware of some decades ago.
Posted by Martine Rothblatt on 05/24 at 11:20 AM
Herve -- I see minduploading via mindfiles + mindware > mindclone as being much easier than nanobot-based readouts of all the mind's contents. We do not have to replicate the neural structure of the brain to recreate a functional equivalent of the human mind. So yes, prior and easier.
Armand -- I agree with you if the challenge is replicating the human brain. But I don't think that is necessary in order to replicate the human mind. The brain is a great way (the only way for now) to get a mind, but I don't believe it will be the only way for much longer. Good enough relational databases, with clever enough software that mimics the way people think and feel, will give us a functional equivalent of a human mind. Hence, I think Kurzweil's estimations are if anything conservative, as it is not necessary to replicate the processing capability of an entire human brain in order to have the functional equivalent of a human mind.
Athena -- The old "you wouldn't kill your original" thought experiment actually doesn't work. My answer is that I'm indifferent to which is killed, although as a pacifist I'd prefer neither be. Of course anyone would honestly be indifferent because the two are equivalent. To say "you wouldn't kill your original" is actually ignoring the argument's premise, namely, that the original and the mindclone are equivalent.
Athena -- It is an interesting question whether a mindclone of myself is part of myself or someone else. I believe they are not distinct entities, but an evolution of a previously singular entity. As Max More points out, the crucial factor is continuity. Because the mindclone is a continuous extension of an original, I would say they are one in the same person, although that same person now has two instantiations.
Giulio -- I agree with you, and I believe the Law will end up making it so as well. When we mindclone ourself, the law will make the actions of one entity the responsibility of both, and the rights of one, the rights of both. In fact, there will be no "both" identities, as there is just one entity albeit in two instantiations. This is a hard idea for us to get used to, but so was "London tonight, New York tomorrow." Or, "it's Mom on the phone." "How could Mom be on the phone, she's in Texas?" People adapt.
When we get into a car, we take a lot more responsibility than our ancestors did when they walked. If we are distracted for a moment, and run someone over, that is a kind of homicide. Worse if we are on cell or drinking. But, each of us makes a decision to drive or not to, and if we drive we must accept the consequences. So it will be with mindcloning. Some people will not want to take the risk. Others will take the risk for the benefits of parallel living. The Law will recognize mindcloned individuals as one legal entity. In case of doubt, cyber-psychologist expert witnesses will decide.
Posted by Giulio Prisco on 05/24 at 11:55 AM
Martine: " It is impossible to not see the first, costly mindclones in 10-20 years, and then a mass-market explosion of them in 20-30 years... I see minduploading via mindfiles + mindware > mindclone as being much easier than nanobot-based readouts of all the mind's contents. We do not have to replicate the neural structure of the brain to recreate a functional equivalent of the human mind. So yes, prior and easier."
In 20-30 years we may well have the hardware and software to run a mindclone at affordable price. What I am more concerned with, is data acquisition. I have my developing mindfile on your CyBeRev project, but I think we need data acquisition rates orders of magnitude higher to develop a workable mindclone (that is, one that could run as sentient software and feel some kind of continuity with the original). Besides the current Q/As, I think we need some high speed, unstructured input channel, for example via the developing commercial neural interface gadgets, with the hope that future software can make sense of this part of the information-
Posted by Lola on 05/27 at 12:48 PM
I am unconvinced by the idea of mindclones possibly because I see the mind as symbolic only on the highest levels and I don't think the high levels work without the lower ones, which I don't think work without a physical connection to a body.
Also most of what I like to do is physical and involves exercising control over my body. Sitting around causes me to be unable to think. In other words, the idea of just a mind clone without a body seems to me to be an illusion.
Posted by Athena Andreadis on 05/27 at 02:18 PM
Those of you who are mindclone adherents seem not to agree whether mindclones are distinct entities or not, and if yes, whether they should be given rights. Similar arguments apply to biological clones. In that instance, it's rather obvious that they are distinct entities. If you believe that mindclones are as good as the original, then they have to be given rights. If you believe that they shouldn't be given rights, you're implicitly admitting that they're not as good as the original.
Of course, the entire argument is just armchair woolgathering, since mindclones are far less likely than manticores or werewolves -- or falling out of an airplane without a parachute and surviving the fall, except in VR, of course, where all fantasies can come true.
Posted by Sorin Gherman on 06/09 at 04:11 PM
You're wrong: unlike chess, the state-of-the-art software for the game of go is not even close to the top human level of play.
(Except on restricted, very small boards, usually used only by beginners or for teaching purposes).
Page 1 of 1 pages
Add your comment here:
|
IEET Blog