Raj Reddy
posted on 02/06/2003 3:20 PM by camilo

I believe Raj Reddy the author of the article is already a superhuman who super intelligence is already way beyond normal. Inded, I believe he represents the first phase of super humans that use unlimited bandwith and memory to increase their knowledge and extra human capabilities

THIS IS A REALLY IMPORTANT ARTICLE & i urge poeple to read it.
posted on 02/15/2003 2:49 PM by jennifergrizzle

In fact Ii've seldom read one so pertinent and intelligent.



It has long deisturbed me that AI should only be developed as a tool.

Re: THIS IS A REALLY IMPORTANT ARTICLE & i urge poeple to read it.
posted on 03/02/2003 3:33 AM by Jeremy

I have been very interested in AI for a long time. But I have done a 180 in my opinion on the dangers it may hold for the human race. Mr. Reddy's article is a fine example of one of the reasons for my turnabout.

Since he began his piece hinting that he will refute Bill Joy's dire predictions, I assume he will give cause for optimism. Sadly, the more I read, as usual, the more concerned I became.

The part that really struck me was when he posed the question "Should we fear superhuman-AIs?" He points out that millions of species co-exist on our planet, and one more should be no worry. He likened the future of ordinary humans to the world animals now live in. Animals, who share the world with a vastly more intelligent creature who has used and abused them, pushed many of their kind to extinction, and isnt done by a long shot.

He goes on with his cheery story by saying these soon-to-come advanced intelligences will only "rarely" come into conflict with us ordinary humans, and when they do, it'll be like the whites stealing America from the Indians.

Oh, what a relief! Here I was worried some disaster may befall humankind!If Mr. Reddy is correct the worst that could happen is a advanced intelligent species will be around to use us, take what they want from us, and dismiss us as if we were a common pest.

Now, I know what some of you are saying, "Hey, this new race is human, but only enhanced with intelligent agents!" That dont exactly brighten my day either. Being human at the core wont make this master-race any more ethical, Hitler wasnt any more compassionate than the Terminator.

Going back to what I said at the beginning of my message, about Raj Reddy's article being an example of why I have turned against AI, so many of the so-called optimistic views of the future I have read are hardly optimistic at all. I cant imagine why anyone would possibly look foward to the future guessed at here, unless of course they were planning on being part of the new master race.

I know my pleas will fall on deaf ears once again, but I sincerely hope people will begin to consider the consequences of greater-than-humans inteligence rather than being transfixed by the silver lining.

Re: Infinite... -- the Power of Recognition
posted on 03/02/2003 9:26 AM by tharsaile

Not to ignore Jeremy's concerns -- you and I have recently debated that topic, Jer -- but the most interesting part of the speech is human Recognition. This was touched on in the penultimate section, just before "Conclusion".

A recent study suggests that very young humans can recognize not only different human faces, but between different chimp and animal faces as well. As the child grows older, he/she will almost certainly lose that ability to the extent that animal faces of the same species will look alike. This is not unlike a child's uncanny ability to distinguish sounds from all the languages of the world - until they grow older and focus on only the language(s) in which they are immersed.

"Infinite Cribnotes?"

In the third (Borg)', Reddy finally gets down to business, stating how intelligent agents can provide a person with information, thus compensating for the apparent biological cap on learning speeds. If I'm not mistaken, this is like swift or instant access to an infinite cribsheet, making the bearer of these agents virtually omniscient. For example, instead of trudging through a six-month Berlitz course in Mandarin, I would be able to understand everything a Chinese-speaker says because the agents would effectively be whispering the answers (translation) in my ear --- is that what the author is getting at?

The question on my mind is this: won't these superintelligent agents REPLACE or at least weaken the human ability of recognition? I'm not suggesting that we will become unthinking blobs just because we have access to so much information (although that has been said about the Internet! :-) ) I'm saying, if foreign agents do all the heavy lifting in the area of recognition, won't our biological ability to recognize become dulled with time (read 'after enough generations')?

Re: Infinite... -- the Power of Recognition
posted on 03/02/2003 3:48 PM by subtillioN

Re: Infinite... -- the Power of Recognition
posted on 03/04/2003 8:34 AM by tharsaile

I've given your response a lot of thought, subtillion, and I'm still not sure how it will all factor into the selection process. Re your last sentence: I think you may very well be right.

Re:
posted on 03/25/2003 11:38 PM by Arnold

This fear is a very common one. As technology improves, we see that our daily activities are starting to be done by machines. And if we looked somewhat into the future, a posible scenario could be that almost everything is done by machines and nothing is left for us.

But on the other hand, if we see the history of computers and how they evolved, we see that at first, almost every code had to be written. And as time passed appeared programming tools that allowed us to save a lot of code, thanks to the use of libraries and procedures that were already written.

With all this tools available, we can make a 1960's program in a few minutes or just seconds, but the fact is that we are not making nowadays the same programs of the 60's, today soft is much more complex.

As artificial intelligence improves there will be lots of tasks done (thought) by machines, but that is what will allow us to think in a much more complex way, allowing us to reach areas that we didn't even imagined some time ago.

Re:
posted on 03/26/2003 8:34 AM by tharsaile

Just for the record, I am very pro-AI. Any reservations I may have are balanced- no, overwhelmed, by anticipations of benefits.

Thanks for the response!

Re: Infinite Memory and Bandwidth: Implications for Artificial Intelligence
posted on 07/21/2003 12:12 AM by olgadorf

Posted in Infinite Memory and Bandwidth: Implications for Artificial Intelligence by Rag Reddy

Before discussing any of the comments in this thread, I would like to offer a few comments on the speech itself. I believe that Dr. Reddy has brought up some very interesting points regarding the future of artificial intelligence, however, I think he may have failed to fully capitalize on ideas that he has presented. For instance, in his idea of 'teleportation', he states that we will be able to 'watch the Super Bowl from the vantage point of a quarterback in the midfield', however he neglects to broach the subject of us feeling the ball, smelling the grass, and hearing the muffled screams of the fans through the helmet. He presents the idea of teleportation as though it were nothing more than watching a large television (he does bring up the idea of remote interaction, but again does not fully extend the idea to it's potential). For someone to see things through the eyes of another person, or even be able to perform surgery from miles away is not necessarily what we would think of as 'teleportation', but when you add manipulation of all of the senses (which is by no means out of the reach of technology within the near future), and we are able to manipulate objects from across the globe, touch taste and smell things as well, that is when it broaches on what we think of when we say 'teleportation'

More importantly, I feel that Dr. Reddy did not describe his theory of 'immortality' as thoroughly as was necessary to convey his point. He states early on in the speech that 'you will be able to capture everything you ever said from the time you are born to the time you die', and then that 'you need to download extragenetic experiences' to a clone of yourself in order to be immortal. What I feel is missing is the link between these two. Because with the combination of thses two statements, Dr. Reddy has given the formulae for immortality. If we can 'download' all of our experiences, and then 'upload' them onto a clone of ourselves every time we feel like jumping our bodies back in time a few decades, then we have not merely approached immortality, we have achieved it.

Of course, there are two basic arguments to this point, the first being the standard 'mind-body' dilemma. If we believe that a human being is more than simply a collection of experiences and neurological information, then of course this wouldn't work, there's no way to download a soul. However, I think that given the nature of the discussion, we can safely exclude this line of reasoning on the grounds that it will serve no real purpose to the discussion and will simply cloud the issue. (I'm not saying you have to give up believing in a soul, I'm simply saying that for the moment, lets pretend that the human mind is nothing more than a bunch of neurological networks, and electrical pulses). Secondly, there is an obvious lack of definition in the terms 'upload' and 'download' that I have used. Less so in the 'download' issue, because we may simply expand on Dr. Reddy's idea of the potential to record every experience in our lives to record all of the information input by all of our senses throughout our lifetime in order to create a 'digital memory' of our lives, thereby effectively downloading our brains onto a hard-drive.

The idea of 'uploading' a person onto a brain, is a little trickier. Of course, if we could simply play back all the sensory information into the clone, we could effectively reproduce all of our memories and cognitive development. However, this has the obvious detriment that by the time the playback was done, the clone would be as old as you, and thus would defeat the entire purpose of the exercise. Fortunately Dr. Reddy has also (albeit indirectly and possibly inadvertently) given us an inclination as to how this could be achieved when he mentions the
'intelligent agents' and how 'thousands of intelligent agents can search and harness the personal knowledge contained in the terabytes of memory on the body and petabytes of memory on the desktop'. In saying this, Dr. Reddy shows us that we need not actually 'upload' all of the information onto the brain, we can simply store it in an intelligence agent, which could then be called upon to retrieve memories, cognitive information, and previously gained knowledge.

One other problem I have with Dr. Reddy's speech is his assumption that 'They [the 'super human race'] will become an virtual nation of the techno-elite who will mainly interact with each other and to a large extent coexist peacefully with the other species on the planet.' Not that I think that this will not be the case initially, I'm fairly certain that should such a 'race' arise, regular humans would lack the cognitive ability to interact prosperously with this group, and very little interaction would occur. However Dr. Reddy doesn't take into account the fact that such a two-tier society would be unstable and would inevitably collapse. This does not have to be an aggressive collapse, in fact I believe that it is much more likely it would simply come in the form of a gradual phasing out of the 'normal' minority, as more and more humans began to use the intelligence agent. (Much like the adoption of any newfound technology or innovation there would be resistance at first, but these luddites would likely eventually either be forced to succumb to the technology, or be left so far behind as to be rendered incapable or participating in any relevant way in society).

I'm not saying this is necessarily a good thing, nor a bad thing. I simply feel that a society such as the one described by Dr. Reddy would not last for any significant amount of time.

I would also like to make some comments concerning the discussion which has ensued from this article. I believe that Jeremy has made some very interesting and valid points, specifically that 'Being human at the core wont make this master-race any more ethical'. In fact I would go so far as to say that is one of the largest concerns with this view of the future. I personally consider myself an optimist and would like to believe that humans in general are ethical and would be able to handle increased technology with foresight and compassion, however I am also a realist and I know that history has proven that is simply not so. I do believe that it is very dangerous endowing humans with too much power, we've shown time and time again that we can't necessarily be trusted. However maybe with increased intelligence, we will finally be able to see the consequences of our own actions.

With respect to the question of whether these advancements will weaken us mentally by doing 'all the heavy lifting for us', I have little doubt that this will be the case, we will become dependant on these machines for our daily lives, and very likely we will not develop certain cognitive processes as strongly as normal. This is no different from some of the 'lost arts' of things like penmanship and calligraphy, give way to newer abilities like typing and font creation. We will develop new skills at the expense of old ones, and the new skills will lead us to bigger and better things. As John Adams said 'I study war and diplomacy so that my children will be free to study commerce and industry and their children will be able to study art and literature'.

Of course the price we pay for this is the dependence on the technology. If it were to suddenly stop working, we would likely have lost the abilities we have sacrificed, much in the same way that I as a student would likely be lost attempting to write an essay without my word processor and search engines for research. But that is the price we pay for technology, and although it is something we must guard against, it is also something that (at least to some extent) is as inevitable as the progress of technology itself.

In conclusion I would like to say that although I believe that the future described by Dr. Reddy is certainly a possibility, it is certainly not the whole story. Any attempt to predict the future using a single technology is destined to fail, and we cannot know exactly how or where this technology will advance, but I for one am eager to find out.

Re: Infinite Memory and Bandwidth: Implications for Artificial Intelligence
posted on 11/04/2003 6:18 PM by Tekrat

Reddy seems to believes that artificial intelligence will not rise from a machine by itself but rather from cybernetic evolution of sorts. That is to say, human beings will use 'intelligent agents' to such a high degree that they essentially become a separate species, a super human race as it were.

'Should we be afraid of the possible emergence of super human race?' is an interesting questioned posed by Raj Reddy. He believes that for the most part we could all live in harmony with maybe minor conflicts but it's the 'minor conflicts' part that worries me. 'On planet Earth, we see millions of species coexisting with each other! I expect the same will happen with this new more powerful version of us!' I highly disagree with Reddy on that statement.

We can see from our daily lives that a superior species neglects the interests of lesser species. Animals all over the world are becoming extinct as we encroach on their territories for the betterment or convenience of our own species. Even within our own species we see the stronger nations doing what they will against weaker nations for their own benefits.

Bill Joy was worried that machines with superior artificial intelligence may take over mankind. At least with pure artificial intelligence, that is, without human involvement, we do not have to worry about emotions and irrationality. With super humans however, we cannot rule that out as they will still have that human aspect in them. I completely agree with Jeremy in that 'Being human at the core won't make this master-race any more ethical'. Time and time again we see that 'absolute power corrupts absolutely' (e.g. Hitler for one).

From a super-human beings standpoint, regular humans would be less productive, yet they would continue to drain resources. As such, they would be looked down upon as a lesser species and as with animals, may be on the road to extinction. If not such a dramatic end then as Reddy says, it 'is likely to create a new form of digital divide.' Does that mean that only certain professions will be open to regular humans, only certain types of services available to them? Oppression has almost always led to revolution in human history and why should this sort of divide lead to different results?

I think that the topic of immortality and how it will affect mankind is hugely simplified by Reddy. It's true, we may be close to reaching the 'capacity' of the brain in terms of terabytes or petabytes but we are far from understanding how our brain stores all this information. Only by understanding how our brain stores all this information and the way it's perceived by us can we even imagine storing our entire lives on a hard drive as Reddy imagines.

This virtual immortality may have huge consequences for how the human race would evolve. If what Reddy says ever becomes a reality, what's to stop people who are facing hard times to just 'backup' their lives, create an embryo and then kill themselves so they can be 'reborn' in a better time? Who would want to live through depressions and wars?

Life sentences for criminals become redundant as there is no fixed definition of life. If we are able to store all our information then we should also be able to delete parts we don't want or need. Criminals may be 'rehabilitated' by cloning them and then downloading their memories except for those negative aspects that made them bad people, similar to the procedure done in the Stallone movie 'Demolition Man'.

Reddy thinks that 'many things will hardly change; our social systems, the food we eat, the clothes we wear and the mating rituals'. Mating, for the main part, is a form of preservation and continuation. A primordial desire for offspring stems from the fact that we want our genetic material to be propagated. Now if cloning and downloading ourselves becomes a normal part of our lives, why choose to have offspring anymore? One would be able to propagate their genetic material and experiences in a more direct manner rather than vicariously through offspring. Imagine a future world where mating no longer occurs, the world is full of clones, genetic diversity has died out and the world is ruled by super human demi-Gods.

Playing God is a dangerous game and the way things are progressing with genetic research and cloning, the future may be very bleak indeed. Ironically, our quest for 'immortality' may be our end.

Re: Infinite Memory and Bandwidth: Implications for Artificial Intelligence
posted on 11/05/2003 11:50 AM by Kenneth_Y

Reddy brings up concerns regarding the emergence of a "super human race" and a new form of digital divide in society. I believe that these things are not concerns of the distant future but issues we are already dealing with today to an extent. Reddy describes a super human race who augments their intelligence through the use of "intelligent agents" - technology that allows humans to acquire and use information more efficiently than biology allows. We already do use "intelligent agents" of sorts, in the form of personal/lap top/palm computers and mobile communication devices. These technologies allow use to share, access and store vast amounts of information and are tools that allow us to be incredibly productive. Reddy explains some of the capabilities of the superhumans as "getting a month's worth of work done in a day, by harnessing and utilizing the power of thousands of intelligent agents." I see that this is already the case. With the use of computers and the internet people now have access to resources that could have requires weeks to have researched offline, goods and services in almost every field everywhere on the globe as well as vast amounts of processing power at relatively low costs. Someone without access to a computer simply would not be able to keep up the same level of productivity as someone who does. This of course has already created a digital divide in society, a divide that is quickly growing as developed nations moves farther and farther ahead in computer advances. There will not be a "new form" of digital divide in the future, rather just a greater digital divide. However this isn't something we are not already dealing with today.

I also hesitate in using the term "super human race", as I certainly don't consider the tech-savvy "computer geeks" of today as a new species of super people (well at least not yet). I do, however, agree that intelligence will naturally emerge from the development and use of computers. I believe there is a blurry line between "user-friendliness" and "intelligence" in software, as we have begun to see in things such as "smart" search engines, databases that "know" what you like and word processors that can finish your sentences. Although I'm sure we can agree that Clippy will not become the next HAL, perhaps it is reasonable to recognize that such "user-friendly" applications already offer some levels of computer intelligence, and that the development of such intelligence is in no way drastic or frightening.

I can agree with Reddy's vision of a future where people become heavily integrated with computers, allowing capabilities that would easily amaze us by today's standards. However, even today we have capabilities that amaze us by yesterday's standards. Reddy's future is not one detached in the distant, it is one we have already begun to realize and is something we have already begun to prepare for.


Ken

Re: Infinite Memory and Bandwidth: Implications for Artificial Intelligence
posted on 11/06/2003 12:19 PM by superxix

After reading Professor Reddy's predictions for the future, I can say with certainty that his ideas have struck a deep chord in my own vision of things to come. His exuberant scientific optimism tries to find a middle ground between the exciting ideas of Ray Kurzweil and the dark times predicted by Hans Moravec.
First of all, I agree with his research-driven funding theory that postulates that investing money to achieve human AI is not feasible for short term market potential. Governments and venture capitalists are not yet ready to invest money when they don't see a quick return on investment. Thus, from the start we can be somewhat satisfied that billions of dollars will not be invested as a whim to build robots with greater human capabilities.
Secondly, I also tend to agree with his analysis of emerging raw computation power and cheap huge memory space and bandwidth. Based on current technological advances and a propensity for the amount of space and bandwidth to increase exponentially, his prediction of a '30 petabyte PC for 100 dollars in 2020' and 'petabytes per second bandwidth' does not seem so far-fetched. The ability to achieve this infinite raw computation power can be seen as the one way to liberate conventional algorithms from the shackles of compensating for resource scarcities.
Nonetheless, Dr. Reddy's provoking ideas come in the form of future predictions in teleportation, time travel and immortality. His form of teleportation is based on the current advances in virtual 3-D reconstruction of real places by having thousand of high definition cameras delivering information to your visual sense. I can imagine that the technology would be allows us to 'teleport' and be in a virtual specific place. However, Dr. Reddy does not elaborate on his teleportation idea of 'bringing the world to us, and bringing us to the world, atoms to bits.' I could see how our atoms could be converted to bits and I cannot believe that they could be projected back into atoms on the football field as ourselves. However, it might be possible to extract the quarterback's senses from his combination of atoms, convert to bits and deliver this information into a virtual world. Thus, anyone can enter the virtual world and experience the same information experienced by the quarterback's senses. As a result, for example, we would be able to 'hypothetically' see through his eyes, smell through his nose and hear through his ears.
Similarly, his idea of time travel needs more elaboration. The idea to convert our atoms into bits for archiving purposes is a good idea for our great-great-great children to be able to see us as if we were still living and conversing with them. However, I do not see how this conversation would occur in real time. Our stored bits can only recreate us in the same way we see ourselves in pictures and videos. In the future, we could 'be' there for our descendants but no real interaction can occur, only scripted synthetic conversations.
Dr. Reddy's idea of immortality is the one that I tend to agree with the most. If we are in fact able to convert atoms to bits, a process that will most certainly require more than just infinite computation power, memory and bandwidth, to pass on our 'extragenetic experiences' to a clone is in fact the one way to achieve immortality. As long as your new clone receives all the information stored in the old brain, our cognitive and memory functioning can pass on indefinitely by restoring your old body with a new one. The question lies in the feasibility and efficiency of 'downloading' and 'uploading' all this information and the overall incorporation with the genetic DNA structure of the new clone. This process also raises more interesting possibilities of immortality. Could we upload our brain information to some other human being, not necessarily our clone? Can we overwrite other people's brain and technically take over their bodies? What if we did not like our old bodies, would we be able to make physical changes to our clones to enhance, for example our physical beauty or strength? In essence, the ethical implications of immortality may present further hurdles that may not be so easy to overcome.
Lastly, the emergence of a super human race may seem like a very bad idea. Although I can imagine the exhilaration of being able to think and perform 1000 times faster and more efficiently than a mere human being, the social implications of a new class of humans can be devastating on our societies. The idea behind an elite social class brings back Adolf Hitler's infamous ideal of an Aryan supreme race. Who will decide which humans will be able to harness the capabilities of the 'thousand personal intelligent agents'? Dr. Reddy's attempt to stave off our fears of this new race is not very effective. The article raises the idea that these super humans will use their capabilities to 'help the poor, the sick and the illiterate.' However, I believe that such super humans, in all their knowledge and wisdom, will be overcome by selfish human traits. Knowledge gives power and power does corrupt people. In fact, I could not see how this new race of humans will be able to coexist peacefully with the normal humans when there is a good possibility that they will try to enslave us for their own benefits.
In conclusion, this article has thrilled me and worried me at the same time. To even consider the idea of immortality, a perpetual state of experiencing life in new cloned bodies is an unspoken desire of any human being. Also, 'time travel' and 'teleportation' are worthy attempts to use different technologies to better our quality of life. Nonetheless, such technological achievements will bring along a multitude of ethical and social dilemmas that may severely restrict their future implementation.

Re: Infinite Memory and Bandwidth: Implications for Artificial Intelligence
posted on 11/10/2003 6:40 PM by chanroy

Dr. Reddy suggested that as memory and bandwidth goes infinity, Artificial Intelligent would expand rapidly. Without doubt as memory and bandwidth become better and cheaper, research in Artificial Intelligent will be beneficial. Actually not only A.I. but any research and any industry would be beneficial.
Dr. Reddy mentioned about how the infinite memory and bandwidth beneficial us, he used the example of teleportation, time travel and immortality can be achieved in some sense because of the advance technology. But I guess he missed the fact that human has 5 senses. I agree with olgadorf that Dr Reddy has actually 'neglects to broach the subject of us feeling the ball, smelling the grass, and hearing the muffled screams of the fans through the helmet.' (olgadorf is not exactly right, smell can be transported through internet, a company called Trisenx, (www.trisenx.com) actually invented a equipment which connected to your computer and simulate the fragrance) Touching is actually an important part of how human sense something. The first well developed sense of a new born baby is touching. Without actually touching the thing we hardly believe the thing is real. It is the same case on the time machine, the simulation of another human is good enough to satisfy the sense of hearing and listening even smelling, it still lack of the touching part. Anyhow such technology would beneficial us in lots of ways.
Reddy emphasis the benefit of petabyte of memory and bandwidth again and again in his speech and pointed out that super human would take us to a much better and create a better world. What about the dark side of this? Infinite memory and bandwidth also means that virus can be spread much faster, and virus can be hidden in the storage system longer and not being found. It is also the ideal environment for hackers. Reddy pointed out that the computer in the future can perform self-repair/ self-healing process, and therefore there is not need to restart the computer, but, what about physical damage to the computer? It seems there is no way that we can retrieve those zeros and ones that we lost, unless we have backup everywhere that we can access to. It is not hard to see that the pay back is the dependence of technology. The blackout which happened few months ago in the North America will be a good example.
I guess it is really hard to achieve the super-human idea. If we can, for sure it is not going to be as soon as 30 years. Dr. Reddy pointed out that the technology is improving in a sense of double exponential. But he never thought about the limitation of the matters. In his speech he said that super human are 'people who can think and act 1000 times faster, using personal intelligent agents.' But the thing is that there is physical limitation. Moreover, Dr. Reddy suggested that the nation states will be irrelevant in the world with the exist or the super human since power and wealth will be concentrate in this segment of population. He further states that super human can coexist with this the rest of the population. Isn't that contradicting? As we all know that natural world in governed by the food chain. Animals with greater power, strength and intelligent will be on the top of the food chain. With the concentration of power and one of the nature of human, selfishness it seems to be impossible to create the 'good world' as what Dr. Reddy said. History already shows us when absolute power is concentrate in a group of people, war and revolution are always the things which will be happened after. How many government and kingdom were overthrown because of the concentration of the power become selfishness.

I am really looking forward to see the 'good world' which Dr. Reddy suggest and I really wish that it will be a great new world when super human come into the world. I wish this segment of super human can really bring the world to a new standard and set the new order of the world instead of a concentration of power which thinks that they are supevior and wanted to eliminate the normal humankind.

Re: Infinite Memory and Bandwidth: Implications for Artificial Intelligence
posted on 03/10/2004 10:22 AM by hawk_y2

Dr. Reddy nullified the Artificial Intelligence concerns raised by Bill Joy, Kurzweil, and Moravec because statistics had shown that there was a lack of investment on IT research. Dr. Reddy then predicted the technology trends in terms of memory and bandwidth. He also suggested the possible benefits from the infinite amount of memory and bandwidth as well as his interpretation of teleportation, time travel, and immortality from 'an information-technology perspective'. As for his argument in regards to the lack of financial resources, it seems that it only applies to U.S. because on March 3, 2004 a Tokyo-based company ZMP has revealed the prototype of a walking robot which presumably will be mass produced by the end of 2004. So, there are other countries that do invest money on AI projects. As for his discussion on teleportation, time travel and immortality, I agree with most of you that Dr. Reddy did not elaborate enough on those topics. However, considering those were just his blue-print-like concepts, I would rather like to discuss his arguments in the section Emergence of 'Artificial Intelligences with Super-Human Capabilities'.

Dr. Reddy did not distinguish the difference between knowledge and information. Dr. Reddy's examples of 'intelligent behavior based on recognition' are simply an accumulation of information as opposed to 'knowledge accumulated by other members of the human race' as he claimed. He believed that by using 'intelligent agents' to search and query from this information pool, individual human capability would be expanded. I strongly disagree with his idea. Just because a person can say or quote something from some source, it does not mean that the person understands what he/she is saying. If what Dr. Reddy said was true, then all the actors would be the most intelligent people in the world if they had played the roles of lawyers, scientists, astronauts, pilots, or doctors because they were able to deliver speeches which involved some in-depth knowledge in that particular profession. Similarly, people would be well-respected scholars as long as they carry an encyclopedia with them.

Having said that Dr. Reddy's idea was an accumulation of information, I would like to introduce a new argument: information overflow. Information overflow is already an existing problem. As the result of the increased popularity of the Internet, there are enough urban legends, hypothetical claims or even false information circulating in the cyberspace, for example the 'Nancy Markle' story. If given infinite bandwidth and memory, it will allow even more junk information to reside. Hence, more resources will be required to enhance the filter functions to sort out such worthless information. On the other hand, even if the collected information is useful, eventually it will be accumulated to a point where it is beyond human capabilities to keep track of, and therefore humans will entirely depend on the so-called 'intelligent agents'. In order words, the 'intelligent agents' will take control over the kind of information we as in human beings should know and should not know.

Finally, Dr. Reddy used the scene in the movie Matrix where Trinity learns how to fly a helicopter in a few seconds as a similar example to show that 'human and the personal agents together would have no such limitation' to learn anything new. That would be possible if human brains were embedded with some computer-processor-like mechanism. He was describing a human body as some sort of math function where the input would be the knowledge retrieved via the 'intelligent agents' and the output would be some instant knowledge. That is to say, what we will need is primarily a compilation of old knowledge that will simply fill up the void in human wisdom because we will no longer need to use our own effort to comprehend anything. Suppose that mechanism is feasible, there is an underlying problem which is the fact that new knowledge or invention will not co-exist. We will live in a world based upon some old models and theories of the past. That means cars will be manufactured the same way from a certain point of time in the future, and one should not expect any new features in the cars or even a new form of vehicles to be invented; otherwise, that mechanism will collapse miserably. Using Dr. Reddy's example again, Trinity probably would have died in a crash if the helicopter was a new model with some different control systems.

Dr. Reddy showed us an alternative vision of Artificial Intelligence in the future. He used a rather more realistic approach to explore the topic by predicting the technology trends in the memory and bandwidth development in the next few decades, and relating their capabilities to Artificial Intelligence development. He suggested a concept of superhuman capabilities with the use of 'intelligent agents'. These agents will supposedly help us to search and query personal knowledge which can be gained through 'recognition memory' instead of complex reasoning. New problems, however, will be created under Dr. Reddy's idea of AI. There will be information overflow which ultimately will give these 'intelligent agents' an opportunity to take control what kind of information humans should and should not know. Also, the society will stop advancing and freeze in a certain period of time because new knowledge will not be discovered by individuals. Because of these probable problems, I do not agree with Dr. Reddy's view on the AI development in the future.

Re: Infinite Memory and Bandwidth: Implications for Artificial Intelligence
posted on 03/10/2004 12:43 PM by rohan909

Hi all

Dr (I can't believe s/he's a doctor) Reddy appears to be a baby-boomer who treats futurism as a form of wish-fulfilment. Anyone who uses The Matrix as an example in a talk about the future should set your bull**it detector a-ringin'.

What Dr Reading tries to call teleportation is simply teleprescence. This is deliberate misinformation in order to get media catch-words into a talk. As is the use of the term "time travel".

Hawk makes some very good statements above in relation to Dr Reading's misunderstanding of what constitues increased intelligence:

Re: Infinite Memory and Bandwidth: Implications for Artificial Intelligence
posted on 03/11/2004 1:04 PM by griffman

Knowledge equals power....... I'll let that one stand as true...

information equals knowledge... the only difference is the matter containing the data. knowledge is contained in a human substrate. information is stored in any other substrate. they are equal.

Knowledge DOES NOT equal intellegence....
intellegence is the ability to process information. its a conversion from incomming data into stored knowledge in the brain/mind. the closer to a 1:1 ratio in the conversion, the higher the intellegence.

Wisdom is the intellegent APPLICATION of knowledge.... to be wise you must have both knowledge as well as the ability to use that knowledge effectivly. you can have all the knowledge in the world but if you are not intellegent it will do you no good and posibly great evil.

knowledge can equal power. power is the ability to govern and direct actions. the more power you have, the more you can direct. you don't have to be intellegent to gain or use power.

absolute power corrupts absolutly. this is a wise stament to tell the unintellegent.

history teaches us that those in great power were not very intellegent. they were clever/lucky/fortunate but not very smart. their consistent failure to follow an intellegent course to success is also well documented.

ethical behavior is an intellegent use of knowledge that governs the behavior of the wise.

can you name a Nobel loriet that was unethical in their scientific practicies?

We must make these distictions when speaking about increasing human intellegence. we are augmenting our ability to reason. and intelegent reasoning will always be ethical. unethical reasoning is a sign of stupidity.

with unlimited access to information and increased intellegence. wisdom becomes much more pervasive. ethical behavior will follow that wisdom.

we will have unlimited access to information before we are able to increase our intellegence. this access is not an increase in intellegence, only information. unintellegent people will run into the barrier of information overload. this can be seen by witnessing managment's use of the internet as oposed to a scientist's use of it.

the question becomes what happens to the stupid people that are currently in power when you suddenly increase their intellegence? this intellegence will lay open the consequences of their actions, past and future. frankly i feel for those that have made unethical descisions to gain or to use their power will be overcome with guilt as their flaws are revealed to them.

there is a major difference between thinking your superior and being superior. people can think they are superior all the time, and they do, at great length. those that most people would agree ARE superior (Einstien is the only reference I care to make here.) would never be caught dead stating their superiority to anyone.

to exploit and abuse animals and the land is very unintellegent. we have to remember that, currently as humans, we are all more or less equal. the indulgences of man are not wanted by those that are above man. there is no reasoning for them. some men SEEN as above man now give us a glimpse of this, as they have forgone these indulgences now.

there will be a devide. and the devide will be voluntary, unlike it is today. there will be many that do not wish to augment their existence and they will have a place in a global society. those that choose to develop will also have a place in society. Each level is dependant on the lower levels for survival and balance. humans have their place and will be content to be human, just as any animal is more or less content being that animal. technology will provide the ability to effectivly jump these levels for those caught in the wrong substrate.

there is a large population of minds that for the most part, have no wish to remain human. though I can't point them out, a number of them pass by this web page. All who come here are deeply interested in ethical behavior of some kind. I ask what you would do if given this opertunity? would you try and control the world? abuse and enslave humans and animals of all kinds? wipe out populations just for kicks? with unethical people, the subject of good behavior never even comes up. this should be a notable sign of a level of intellegence. increasing an unethical person's intellegence must first past your, as an ethical person, level of intellegence before they reach greater than your high human level.

If such a group as we here are allowed to become more than human, they will have little need for things like money and power. the desire for such things would not consume them. they would not hoard more than they require because it is unwarented. interactions with humans would be on an as needed basis and would be met with compasion and respect. it is the intellegent way to live and think.

almost all our fears are based on the fact that the majority of people in the world are selfish and stupid. the two are never far apart. we are talking about the ability to remove that stupidity. the selfish behavior would die away just as fast.

what evil can be done with power in the hands of the wise?

griffman

Re: Infinite Memory and Bandwidth: Implications for Artificial Intelligence
posted on 03/11/2005 4:27 PM by nicosk

Dr. Raj Reddy has an interesting perspective into the future. However, I am inclined to not agree with all of them.
His basic premise is that the money spent on developing artificial intelligence that is up to par with human intelligence is not enough. Instead, Dr. Reddy predicts that technological advancements will bring us closer to intelligence agents that will be made available to humans.
My perspective into the future is quite different. I do not believe that advancements made on technology available today will bring us closer to a more 'intelligent' future. I am waiting for that breakthrough technology that will make everyone revise all their predictions. Perhaps this technology will be grid-based Cell networks, perhaps this technology will never materialise.
Nevertheless, let's examine Dr. Reddy arguments. It is quite possible that we will be able to store and transport the volume of information that Dr. Reddy proposes can be used to store all our experiences. However, I do not believe that a serial computational model suffices to process this information. At least not a computer based on technology available today. If the information cannot be processed in real time (so that memory is not a non-issue as Dr. Reddy proposes) then this information is unusable.
But let's consider a world where this processing power is available. This world has stored information about everyone's life experiences and is able to process them efficiently. We must now consider how this information has been captured and who controls it.
Maybe it was you. Maybe you had the technology and more importantly the financial resources to capture all your life and store them in something other than your brain. This is highly unlikely. Even if the technology existed, the financial obstacle will not easily be overcome. This argument will hold for most people in the world.
This however does not mean that your information is not stored somewhere. What if a marketing company has stored all your life in its vast database and sells this information to any interested companies. This is not a far fetched idea even today ('Do you have a insert-superstore-name-here card?', Doubleclick, etc.) The only difference is the amount of information stored. The problem however arises on how this information was captured.
This seems to be the biggest obstacle. But perhaps not. Instead of a marketing company, imagine that the information is stored by your government. How would the government obtain this information? Well, the government could legislate this concept into existence. If technology exists that can transport and store your life, then maybe there is technology that captures this information indirectly (for lack of a better term, a sensory sensor) or directly (a microcontroller or microprocessor embedded into the human brain that tracks the entire nervous system).
What would the government do with this information? Automate the taxation process, the voting process, just about anything one can imagine. In fact, the magic number here is not 1729, but rather 1984.
I said previously that this technology would be too expensive for people to use. How could the government handle this cost? One way would be to sell this information to companies. If privacy laws are applicable in this world, the amount and type of information sold could be limited.
But perhaps I got it all wrong. Maybe all of this information is accessible at a low cost. In this case governments and companies could still have access to your life, but so would you. What would you do with this information? For one thing, you could fallback to the storage device for your memories. No more 'What did I eat yesterday?' or asking your mother about something that happened to you when you were two years old. But surely there is better use for your life experiences. You will no longer have to remember what you have learnt (if learning is a concept in use at this point).
If we can store our life onto a storage device, why not store someone else's experiences as well? This way you will know how it feels like to be an Olympic athlete or how weightlessness feels. Why stop with real experiences? We could add experiences conceived by a human or randomly generated by a computer. Imagine ice skating on the surface on a star, going past the event horizon or walking on your tongue.
If all of this seems plausible to you, then this is where our trains of thoughts will follow different tracks. If all of this is possible, then we have already discovered how to store and retrieve thoughts. This means we have decoded our nervous system. Once this has been achieved, then creating a robot with artificial intelligence is possible, something Dr. Reddy says would be financially very demanding. Even though he has argued this statement earlier, and proposed intelligence agents as an alternative, his ideas converged with Bill Joy's ideas. If we can decode, store and retrieve thoughts, memories, experiences, then a robot could be programmed to use this stored information, examine their surroundings and establish patterns with the stored information and duplicate the action taken by humans. If the robot can project situations then it has become intelligent and perhaps even creative.
Let's dismiss my last comment and examine the three perspectives that Dr. Reddy sees for the future of Artificial Intelligence: teleportation, time travel and immortality.
It's already been mentioned that the concept of teleportation as described by Dr. Reddy is more closely related to telepresence. This concept has already been materialised. Books, radio, television and so many other media try to immerse us into a different world. In these worlds, you are the quarterback and you did win the Super Bowl. Our brain performs teleportation every time you wish you were in someone else's shoes. Dr. Reddy's is a substitute that seems to have been thrown around as a buzzword rather than an Artificial Intelligence goal. It will bring artificial environments closer to reality, perhaps even make them 'feel real', but this is not teleportation.
The next argument is rather more interesting, time travel. Dr. Reddy's time travel is not equivalent to the definition of time travel, but still presents an interesting idea. One could store experiences that have already occurred and re-experience them in the future. Nevertheless this is equivalent to telepresence/teleportation. What Dr. Reddy imagines is the capability of having an imaginary conversation with Einstein. However, this presupposes the existence of some sort of artificial intelligence to control Einstein's responses. The entire premise of Dr. Reddy's ideas is that such artificial intelligence is too expensive to be developed. In order for Dr. Reddy's time travel to work, we just need to link this artificial intelligence with a much larger data pool, i.e. Einstein's life experiences.
As far as time travel into the future, well this is just the extension to Dr. Reddy's teleportation. This teleportation was limited to 'portals' that have been generated based on captured data. Time travel into the future is equivalent to Dr. Reddy's teleportation to portals that have been created rather than captured. In fact the only relation with the future is that it is 'presumed' to be in the future. Take as an example a simple situation. You wake up on Monday, but everyone you meet says it is Tuesday (the coming Tuesday not the one that has just passed). You are now in the future, a much more real future than the one Dr. Reddy proposes.
The third concept that Dr. Reddy proposes is immortality. This involves transferring all our extra-genetic experiences to the brain of one of our clones. This will be achieved using a simulated learning environment. My mind cannot conceive how this process will occur. Will the clone not experience certain situations in the process of being born? Will these be erased, or will they be replaced by our own similar experiences? Once born, will we prevent them from experiencing their own life? How will the experience be infused? If we infuse them with all of our life experiences, at what ratio will this occur? If it is a faster rate how will they be the same when they will have the body of a ten year old but the experiences of a sixty year old that knows what liquor tastes like, what sex is, who has two children, etc.? Will it experience the same diseases that we have? How will we sustain the clone while infusing the memories? What happens outside the learning environment? Will these experiences somehow not influence the clone? Will the clone realise what process they are involved in? At the end of the learning experience, will they realise who they are or will they think of it as a television show and expect to see the next episode? How will they react to our death? I'm sorry but I have nothing but questions about this concept. I am inclined to say though that immortality as Dr. Reddy envisions it is more likely if we replace the clone with a robot which has enough artificial intelligence to realise that it will undergo a learning process during which it will experience a person's life experiences.
Dr. Reddy then talks about humans that have access to super-human intelligence. This intelligence is made available through intelligence agents both physical and in cyperspace. These agents would recall experiences and knowledge and make them available to us.
Dr. Reddy claims that human intelligence is limited by the knowledge that can be acquired during one's lifetime. I will agree with him, but I will try to prove that this limit is to an extent virtual. We can transfer knowledge that we have acquired to someone else, so that they can acquire knowledge that we were not able to or for some reason did not. We use libraries to collect all this information so that even though we cannot know everything we have access to all knowledge that has been made available to us. And as people 'create' more knowledge, the limit stops being an impediment.
What Dr. Reddy envisions is a data pool which we can query for information we do not have. At least I hope that this is what he envisions since his personal agents could be robots. Assuming that the personal agents are in fact data pools (or possibly an interface for querying these data pools), once again we return to the impediment of how this data is collected and who controls it. We have reached a full circle.
It was nice to see some thought being put into commenting Dr. Reddy's article by the other commentators. Brian Harrington's comment about not experiencing everything during teleportation and Tekrat's question of whether or not mating will be needed in this world, were especially striking.
Dr. Reddy has definitely put a lot of thought into this topic as opposed to me. However, I have tried to fill in the holes (and make new ones) in this thread. It's not science fiction, it won't be in the future, it's already here. We have amassed large data pools (google, msn encarta, public libraries, private libraries, etc.) and we have restricted access to them (northernlightsearch, websites using robots.txt). As we enlarge these data pools and process them to create information pools, we will be one step closer to creating knowledge pools and therefore one step closer to accessing intelligence (artificial or not). If you count how many times I typed in 'perhaps' and 'maybe' you'll realise that there is no certainty in this topic. Only time will tell.

Re: Infinite Memory and Bandwidth: Implications for Artificial Intelligence
posted on 03/11/2005 5:34 PM by grantcc

Re: Infinite Memory and Bandwidth: Implications for Artificial Intelligence
posted on 07/14/2005 5:46 PM by whitemage

Dr Reddy's observation that businesses and governments seem to prefer ventures with near immediate capitalization is very astute in that it coincides with human nature. One of our characteristics of being human is the need for instant gratification and wealth. With this in mind, I do not dispute the emergence of technologies that mimic what Reddy calls 'teleportation, time travel, immortality and superhuman capabilities'. What I do dispute is whether humanity will evolve so as to not abuse these new technologies.
In the case of immortality where memories can be downloaded and then installed into a clone or other body, what safeguards are there to prevent the modifications of this information? How can this information be stored so as to become useful when the original host has died? Preservation of digital information is a problem faced today and will be faced in one way or another in the future. With the frequency of paradigm shifts accelerating in the future [Avoiding Technological QuickSand: Finding a Viable Technical Foundation for Digital Preservation. Council on Library and Information Resources, pg 4], it may become extremely difficult to save memories. The modification of memory can involve both input of new fallacious information, or removal of said information. This would make persecution of criminals extremely difficult as they then have plausible denial. Such situations occur often in science fiction movies such as 'Paycheck' or episodes of 'Stargate SG-1'.
The nature of this form of immortality will raise some interesting questions regarding religion. Many religions believe in the idea of a soul or divine spark. If a person dies and his or her memories are saved then successfully implanted, is this not the same as creating new life? A phrase I like to use is, 'I wish life had save points. Then I could just reset if I screwed up'. This mentality would be a problem if we could simply clone and upload a new body each time our lives takes a turn for the worse.

The merging of technology and artificial intelligence to create 'superhuman capability' has the most social and ethical issues out of all the points presented by Reddy. With the current digital divide between have and have-nots, only those who can afford or have access to facilities can initially partake of the enhancements. This will facilitate racism against those who cannot afford the same privileges. Historical precedence can be seen in the treatment of African Americans by Caucasian Americans and the Jews by the Nazi. In addition, as Griffman has said, there will be resistance by those who want to remain 'human'. Again there has been historical precedence in the form of the Luddites [http://en.wikipedia.org/wiki/Luddite]. Reddy has hinted that this may be possible, yet indicates that 'On the rare occasions when there is a conflict, as it happened when the Native Americans confronted the Settlers, it will be an uneven contest!'
Reddy makes fascinating arguments about atoms to bits. Much of the technology in science fiction can be reinterpreted so it can be available within our lifetimes. The pace at which technology is progressing is staggering. However, Reddy blithely assumes that society will learn to accommodate the inevitability of technology. It was indicated that this will change the way 'we live, work, and govern ourselves'. While this may be true, the social impact of these technologies must be carefully examined to prevent misuse. Otherwise, the first faction to develop the technology will become the 'Settlers' and the rest of us will be remitted to the role of 'Native Americans'.

Re: Infinite Memory and Bandwidth: Implications for Artificial Intelligence
posted on 11/28/2005 5:08 PM by mo.sadoghi

Teleportation:
________________________________________
' Every element in the vector space of Rn has n-dimension and is presented using n-axioms ('Vector Space'). Earth belongs to R3 and every element that can be observed by a physical entity from within this space is also represented in 3-dimension. We are living in a 3D world, therefore if an incident exists in which we can observe using our five physical senses then that incident is also a physical entity.
' Human beings are collections of cells consisting of billions of atoms. Assuming that there exists a mapping algorithm from an atom to a bit then every atom of human body can be transformed to a bit, which resembles a virtual version of a human being having the same set of human characteristics. Human behaviors are a sequence of reactions that can be felt in the physical world and they are a part of this world, therefore they must be represented using physical objects.
' Therefore, the virtual human carrying the same set of properties stored and presented using a sequence of 1's & 0's should react and behave identical as a human made of atoms, given they are both created and simulated in the similar environment.

--> Hence, teleportation is realistic.

Time Travel:
________________________________________
' A computer program can be stored, ran, and executed infinitely many times. If teleportation is possible it will imply existence of a virtual human. Thus, one could potentially activate and run a virtual human on a simulator at any desired time, which allows interaction with a virtual person in the past, future, or both.
' If a snap shot of virtual human, H, is taken at the time t0,-2- then we could only interact with the virtual human stating at time t0. Our interaction with H at time t0 ' t1 is not valid because H's physical world has changed significantly from t0 to t1 and these changes are not reflected in computer image of H. In the interval from t0 ' t1, at the very high level the earth has revolved with respect to itself and the sun and other solar planets and the Milky Way has also changed its position relative to other galaxies. Therefore we could only speak to H at instance t0. An instance is infinitely small which implies we can not actually interact with H at t0.
' Technically we could interact with H from t0 ' t1, if we are able to construct a virtual model of the universe and simulate all the possible changes from t0 ' t1. In order to construct a virtual universe we have to map every atom of universe to a bit. The universe consists of infinitely many atoms and the transformation requires infinite number of bits and a bit is also bigger than an atom (since bit is made of atoms, too) therefore this mapping is not feasible due to the lack of space and resources.

--> We can not construct a virtual universe; hence we can only talk to H at instance t0 which is not feasible, either.

' Now, hypothetically imagine we could construct a virtual universe, this would allow us to know exactly how person H undergoes from t0 ' tn, where H terminates/dies at tn, when it is simulated on the virtual universe.
' But, now the question is that if H lived 200 years prior to us, is it possible to interact with H? Our existence 200 years ago could have potentially affected the behavior of H.-3- Thus, we could theoretically speak to H in the virtual universe but H's response to an external entity which is not a part of the virtual system is invalid, since we did not exist when H existed. Moreover, we are contradicting the very basis of our construction of Teleportation. At the time of H we were not part of the virtual vector space indicating that the effect of the external source was not reflected in the creation of H. Therefore, an external source could not interact with the vector space model making our interaction with H invalid.
' Hence, we could never truly interact with H who lived before us, so we cannot travel back in time; the same argument could be used to prove we can not travel into the future.

--> A true time travel is not possible.-4-

Immortality:
________________________________________
Reddy's analogy of immortality is questionable. The author believes that immortality is the ability to transfer knowledge and experience from one human being to another. Theoretically, conveying information from one brain to another is very similar to carefully studying another person and learning and exploring everything about that person, but yet you are completely different individuals. One can study Newton's life and learn all of his knowledge, but he/she does not become Newton nor does it make Newton an immortal person. It's possible to simulate and transfer physical entity from one object to another (experience and knowledge), but we can not simulate life. There is a difference between life and learning. Life is not a physical entity and can not be presented using atoms, therefore immortality is not feasible. However, sharing and transferring knowledge is definitely possible. The immortality explained in this article is simply an advanced way of passing on information using the latest technological innovations.


Footntoe:
________________________________________
-2- At time t0, every atom of H was transformed to a bit in order to construct a virtual human.
-3- Every object in the Universe attracts every other object with a force directed along the line of centers of mass for the two objects. This force is proportional to the product of their masses and inversely proportional to the square of the separation between the centers of mass of the two objects ('Gravity'). Therefore, even based on laws of Newton our existence 200 years ago could have affected H.
-4- However, this does not proof that one can not watch and observe simulation of the virtual universe. He/she can forward and/or rewind the virtual universe but simply can not be a part of nor interact with it.

Re: Infinite Memory and Bandwidth: Implications for Artificial Intelligence
posted on 11/28/2005 5:47 PM by eldras

wicked!

BTW
teleportation's already been achieved, physically a few years back (Google0 using lasers.

Re: Infinite Memory and Bandwidth: Implications for Artificial Intelligence
posted on 03/15/2006 11:15 PM by vectrln

Dr. Reddy provides an interesting view of the future regarding how the implications of having infinite bandwidth and infinite memory available. I agree with his view that advances in technology will widen what is possible with technology to never before seen milestones (like possible immortality). Specifically, the idea of 'experiencing' events with either in present or past (though in a virtual manner) can serve as a beneficial educational tool which would be more engaging than simply reading about the same event in a book.
Since most people have discussed Dr. Reddy's other points on Time Travel, Teleportion, and Immortality, I would like to focus my discussion on his claim that a race of 'super humans' would arise. I agree that technology could evolve to the point where humans could use 'intelligent agents'. However, his thought that the use of 'intelligent agents' could eventually lead to an enhanced race of humans capable of thinking many times faster, I disagree with. At any rate, such devices could lead to a more privileged race with better tools, but not a super human race.
For starters, such intelligent devices are still just sophisticated tools. A human may gain advantage by using such tools, and would require sufficient intelligence to use them, but not a lot more to warrant calling the users of these tools 'super human'. As stated by Roy Chan, although software may evolve to a state where it can be self-repairing, it can't do much to account for the physical damage of a device and may even be vulnerable to viruses. How can the software tell when its underlying hardware has been damaged? How can a human tell if the data being given to him/her is not actually corrupt data resulting from a virus? At some point human intervention would be required to realize and correct an error. This would require someone familiar with these tools like a mechanic is familiar with a car to fix a problem.
Another issue is power and heat. These devices are only beneficial for as long as they have power. It was stated by Kenneth Y that 'intelligent agents' already exist today that enhance our productivity such as PDAs and Laptops. But intensive use of these portable devices available today tax the power supplies of these devices, resulting only in several hours of usage before their battery dies. Assuming that these 'intelligent agents' of the future are more powerful and efficient, they would still require power to run, especially if they are constantly processing data on behalf of their user. Unless an efficient power source becomes available, these intelligent agents would need to be recharged constantly, which would limit their usage. Furthermore, assuming infinite bandwidth and infinite memory, the amount of information being processed by these devices would generate a staggering amount of heat that would need to be dissipated in an efficient manner that is unobtrusive to the user wearing these devices. Although the future may provide solutions to the problem of power and heat, today there is no escaping the physical consequences of using faster processors able to handle and analyze large amounts of information.
Reddy also describes information being collected from multiple sources by these intelligent agents. But what about the quality of the information being received? As stated by hawk_y2 infinite bandwidth increases the possibility of more junk floating around the 'collective network' of humans. Junk is not the only factor, there is also the human factor of bias and exaggeration that can skew the data being retrieved. How can a software program knowingly be objective about information it collects in order to give an adequate analysis to its user, especially if the user is counting on accurate information? If a human needs to think about the information it is receiving, rather than trust in the intelligent devices themselves, the more the devices become relegated to the position of tools for information. Such tools would only aid in data collection and not a person's actual thinking. In order to process at the speed 1000 times faster than humans, intelligent devices would need to be able and think on behalf of the human 'hosts' that they reside on and the result immediately acted upon by their human host.
Dr. Reddy provides some interesting insights on the future, but I would argue that the 'intelligent agents' he describes are merely tools that would increase the productivity of its users. As stated by others in the discussion such as hawk_y2, mere accumulation of data by these tools would not increase the intelligence of the person using them.

Re: Infinite Memory and Bandwidth: Implications for Artificial Intelligence
posted on 03/20/2006 12:02 PM by atsang

I've always been an avid support for Artificial Intelligence just because I'm intrigued with what the limits of technologies would be. This article I thought was a rather interesting because it focused on the possibilities when limits were lifted on storage and bandwidth constraints.