http://www.neurdon.com/wp-content/uploads/2009/04/the-future-of-ai-300x258.jpg
300
258
Ray Kurzweil has been long known as one of the main "futurologists" advocating the loss of intellectual supremacy by humankind to machines in the near future...actually, in the next few decades. Such a topic cannot be left without comments by Neurdon!
In summary, Ray claims that "...by around 2020 a $1,000 personal computer will have the processing power of the human brain—20 million billion calculations per second. [...] By 2055, $1,000 worth of computing will equal the processing power of all human brains on Earth (of course, I may be off by a year or two)."
Ray extrapolates these figures by looking at how various technologies have evolved at an exponential rate in the past centuries. Even if Moore's Law is slowing down, "...new computer architectures will continue the exponential growth of computing. For example, computing cubes are already being designed that will provide thousands of layers of circuits, not just one as in today’s computer chips. Other technologies that promise orders-of-magnitude increases in computing density include nanotube circuits built from carbon atoms, optical computing, crystalline computing and molecular computing."
There is little or no doubt that increase in computational power will soon be available to simulate large-scale neural systems of the complexity (or, at least with a similar number of neurons) of a small mammal, or the human brain, in the next few decades. The SyNAPSE project is an example of such efforts to overcome current chip limitations. Also, there is little doubt that "...once computers achieve a level of intelligence comparable to that of humans, they will necessarily soar past it."
I agree with Ray's main conclusions, despite we can all be skeptical about the rate at which this is going to happen (we can be off by 10, even 100 years, but this does not change the main trajectory of where technology is heading). What is more challenging is to understand what this technology, and new form of intelligence, will look like in, say, 50-100 years from now.
While the focus of attention is mostly on when we would be able to "replicate" mammalian intelligence, starting from animal senses to motor action, passing trough intention and planning, new forms of intelligence that might emerge when adaptive systems are hooked up to "non-conventional" senses are far more fascinating, and their evolution trajectory more difficult to grasp. Dharmendra Modha, leader of the IBM SyNAPSE team, hints to this "extended" form of artificial intelligence, remarking how "non biological sensors, sensors such as monitoring the forest, sensors such as monitoring the ocean, sensors such as monitoring people, animals, organization, homes, cars..." can be mined to "....extract patterns, large scale invariant patterns from the sensory overload, and to act and respond to these data."
It is not immediately clear that the "cognitive architecture" that will support completely novel kind of sensors and applications will have much in common with what neuroscience is studying today.
Below a video of Ray Kurzweil going over the main concepts of the singularity.
[...] Max’s recent post brings up the issue of Ray Kurzweil, a polarizing figure if there ever was one. First, I try to take his musings with a truckload of salt, but his proclamations about the progress of neuroscience seem to go so far beyond the pale that its shade has surpassed the visible light spectrum. He seems to completely trivialize the daunting task set before neuroscience: create a biologically and mathematically precise functional characterization of the human brain. In other words, how does it do what it does with what it has? It seems like Ray is fixated, naturally, on the exponential growth of information technology and its implications for the field. I’ll summarize his thesis, based on the 7 minute video linked by Max, thusly: computers will become insanely powerful, along with ways to measure the brain, and so we’ll have a brain simulation by 2020, or 2029, or sometime in the next fifty years. [...]
ray kurzweil has gone form being a genius engine. er/programmer to becoming a a demogogue.
one can speculate why he left inventing for politics, but presumable he did so for the same reason most engineers who choose to go to politics do so, because he recognizes that creativity and inventiveness come from society at large and that affecting and influencing the entire organism in this direction may be perhaps more interesting, more productive for science, and more fun than sitting around in a lab all day inventing things in isolation or with a small team.
ray is leveraging his prominence and personal fortune to gain influence as a policy maker and promoter (albiet so far on a mostly private level). he is , much like al gore, persuading the public of an imminent event of importance to everyone coming in our future. the singularity/global warming whatever……..
the bottom line is that you can’t predict weather, nor can you predict technology. those who do so are trying to get attention to further their goal. for ray , it’s the goal of getting society to care more about investing time and money and effort into techonology. i think this is a worth goal. for al gore, it’s getting people to care about conerving our environement and eliminating pollution.
while both goals are worthy, both men are approaching these goals in a faulty way. you cannot convince people to approach a goal with a promise of something untrue, because in the end, the truth is discovered, and in their dissolution, soceity at large forgets the worthiness of the ultimate goal that is being pursued.
what if ray is wrong and there is no singularity…..and then all of his institutes and investors walk away. it would be a shame , because they are probably working on worthy projects , and yet, the promise of the golden singularity does not need to be the thing that keeps people working towards the positive ends of building intelligent machines.
personally, it think ray is 100% wrong. i don’t think we will have machines as intelligent as animals for quite some time. to be as intelligent as animals, they have to have sensory systems , emotional systems, and many other systems all working in concert with some all powerful distributed processosor. i just don’t see it happening. what seems to be more likely is that machines will continue to expand their sphere of activity towards those marginal areas where they can be used most successfully in replacing human beings. i.e. where robotics can be combined with intelligent machine computing to develop ‘expert machines’ that can learn with experience to complete basic tasks with an acceptable margin of error.
i can think of a few….dogwalkers, coal miners, subway clerks,etc….
i’d say that is a mmuch more conservative prediction than ray kurzweil’s singularity.
Hi Zeev,
thanks for the reply! You say above that “personally, it think ray is 100% wrong. i don’t think we will have machines as intelligent as animals for quite some time.”
When do YOU think this will happen
Max
Just killing some in between class time on Digg and I found your post . Not typically what I like to learn about, but it was absolutely worth my time. Thanks.