Matt,
First, it is not clear "people are free to decide what makes pain "real"," at least subjectively real. If I zap you will a horrible electric shock of the type Sadam Hussein might have used when he was the chief interrogator/torturer of Iraq's Baathist party, it is not clear exactly how much freedom you would to decide how subjectively "real" the resulting pain would seem to you --- that is, unless you had a level of mental control far beyond that of most humans. You indicate we currently don't know the degree of consciousness or pain that would be suffered by a certain organism with 302 neurons. I agree. Our understanding of the physical correlates of consciousness is still relatively limited, but it is rapidly increasing. I think it is probable that consciousness comes in various decrees, and it is possible that all of physical reality has a form of consciousness, just one that lacks many of he attributes of a human consciousness. A 302 neuron nervous system may have a type of consciousness, but it is my belief it would be one so much less rich and complex than that supported by the 100,000,000,000 neurons of a human brain that it is not only different in degree but also extremely different in kind. I understand I am making a statement based on belief when I predict we will make great strides in understanding the physical correlates of consciousness in the coming fifty years. But there are already a number of studies shedding light on that subject. If we have anything close to the advances in brain scanning and brain science that Kurzweil predicts 1, we should come to understand the correlates of consciousness quite well --- so well, in fact, that we should have pretty good, although not necessarily complete, explanations for the various facets of the Chalmers' hard problem of consciousness. That is, we will come to understand that consciousness is created largely or entirely by computations in physical reality, and we will develop a fairly broad understanding of what type of physical computations yield what types of subjective conscious experience. With this knowledge we would be better able to understand the physical correlates of conscious pain, and, thus, better estimate the probability that various humans, animals, or machines will suffer something like pain under what circumstance. The hard problem of consciousness is based on the assumption --- or at least the question whether --- consciousness has aspects that are separate from the physical world. As we increasingly learn more about the physical correlates of consciousness, I think the scope of the hard problem will increasingly diminish. Yes, there are things about consciousness that we cannot clearly define in terms of physical computations at this point in time, but it is not clear that will always be the case. Just as life is created to various degrees of complexity out of bio-chemical computations, I think human consciousness will be shown to be created to various degrees of complexity out of neurological computations. It is conceivable that the properties of other levels or reality will be required so explain some physical correlates of consciousness, such as such as quantum entanglement or quantum weirdness. I think future study will probably tell us if this is necessary. But ultimately there will always be limits to our knowledge. We have no ultimate way of knowing with total certainty that our perceptions of reality are anything other than an illusion. I agree with Richard's paper when it points out the often repeated statement that our subjective experiences are the most "real" things we have. But just because they are subjective to us now, does not necessarily mean that they are largely beyond the scope of human and AGI assisted science. Ed Porter 1. Kurzweil has claimed we will be able to so accurately scan and model an individual human mind that we will be able to create a virtually exact duplicate of it, including is consciousness, its memories, its passions, etc. I personally think that is unlikely within 50 years. But I think that the combination of brain science and AGI will allow us to understand the mysteries of the hard problem of consciousness much better in fifty years than we do today. -----Original Message----- From: Matt Mahoney [mailto:[EMAIL PROTECTED] Sent: Monday, November 17, 2008 12:44 PM To: [email protected] Subject: Re: FW: [agi] A paper that actually does solve the problem of consciousness--correction --- On Mon, 11/17/08, Ed Porter <[EMAIL PROTECTED]> wrote: > For example, in > fifty years, I think it is quite possible we will be able to say with some > confidence if certain machine intelligences we design are conscious nor not, > and whether their pain is as real as the pain of another type of animal, such > as chimpanzee, dog, bird, reptile, fly, or amoeba . No it won't, because people are free to decide what makes pain "real". The question is not resolved for simple systems which are completely understood, for example, the 302 neuron nervous system of C. elegans. If it can be trained by reinforcement learning, it that "real" pain? What about autobliss? It learns to avoid negative reinforcement and it says "ouch". Do you really think that if we build AGI in the likeness of a human mind, and stick it with a pin and it says "ouch", that we will finally have an answer to the question of whether machines have a consciousness? And there is no reason to believe the question will be easier in the future. 100 years ago there was little controversy over animal rights, euthanasia, abortion, or capital punishment. Do you think that the addition of intelligent robots will make the boundary between human and non-human any sharper? -- Matt Mahoney, [EMAIL PROTECTED] ------------------------------------------- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?&; Powered by Listbox: http://www.listbox.com ------------------------------------------- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244&id_secret=120640061-aded06 Powered by Listbox: http://www.listbox.com
