Kurzweil's rebuttal to Paul Allen
technologyreview.com
Kurzweil's rebuttal to Paul Allen
1–10 of 123 posts
Re: Kurzweil's rebuttal to Paul Allen
#2Kurzweil makes some good points here. He doesn't address each and every criticism with his view of AI progress, but he does a good job calling out Paul Allen on not doing the homework.
Re: Kurzweil's rebuttal to Paul Allen
#3Kurzweil makes some good points here. He doesn't address each and every criticism with his view of AI progress, but he does a good job calling out Paul Allen on not doing the homework.
Calling out someone for not doing their homework seems somewhere in the DH1-2 range:
http://www.paulgraham.com/disagree.html
There are some good points in Kurzweil's response, but the ones about Paul Allen are definitely not them.
I think his best point was about extrapolating function from individual cells or structures, without needing to understand every single cell or structure individually.
Re: Kurzweil's rebuttal to Paul Allen
#4Science progress is not just a question of intelligence.
Therefore, even if computers become more intelligent than humans, it is doubtful a "singularity" will occur.
Re: Kurzweil's rebuttal to Paul Allen
#5The Singularity concept strikes me as a sort of wishful thinking. Technology advancing so fast that we no longer can control or understand it? Yeah, that already happened to my parent's generation with AOL, yet here I am texting this on my iPhone. New generations understand intuitively what the previous generation understood theoretically.
Even so, I fully expect memristors to deliver strong AI.
Re: Kurzweil's rebuttal to Paul Allen
#6The Singularity concept strikes me as a sort of wishful thinking. Technology advancing so fast that we no longer can control or understand it? Yeah, that already happened to my parent's generation with AOL, yet here I am texting this on my iPhone. New generations understand intuitively what the previous generation understood theoretically. Even so, I fully expect memristors to deliver strong AI.
> Even so, I fully expect memristors to deliver strong AI.
Why would they?
Is strong AI a function of storage capacity or speed?
An AI running at 1/100th of what a future AI may be capable of is still an AI and I can't see how a mere improvement of a couple of orders of magnitude would do what decades of Moore's law have failed to do so far.
If strong AI was just a matter of speed then we could theoretically take any of the large clusters available today and run it at an appreciable fraction of its normal speed leading at a minimum to a validation of the fact that it is indeed a strong AI that's been created.
The barrier seems to be more that we don't know how to go about building one from a software perspective than that we wouldn't have the capability to design the hardware.
So how would an advance in hardware suddenly fix that?
Re: Kurzweil's rebuttal to Paul Allen
#7Kurzweil makes some good points here. He doesn't address each and every criticism with his view of AI progress, but he does a good job calling out Paul Allen on not doing the homework.
Calling out someone for not doing their homework seems somewhere in the DH1-2 range: http://www.paulgraham.com/disagree.html There are some good points in Kurzweil's response, but the ones about Paul Allen are definitely not them. I think his best point was about extrapolating function from individual cells or structures, without needing to understand every single cell or structure individually.
I disagree. When one "addresses" a long-standing, well-thought out argument with off-the-cuff snap statement or something only slightly more thought-out, it is perfectly permissible to call them out as not being serious. It's like going into a serious religion debate with "Evil exists, therefore God does not. Ah-ha, you are defeated!", or "God must exist because something must have started it all, so there!" as if in the thousands of years of years the debate has been raging on nobody has ever thought of those things, or addressed them at length, in both directions.
If you're going to debate the singularity here, maybe you can get by with just stating "I don't believe it's possible" without citing any logic, which as of this writing there's at least two people in this comment set already who have simply stated that without defense, but if you're going to debate one of the leaders of the field it would help if you would at least grant your opponent the courtesy of thinking that just maybe over the course of the decades he's been thinking about this, the obvious objections that you thought up in five seconds just might have been addressed at some point. You may not think they've been adequately or correctly addressed, but don't pretend they haven't been addressed at all.
Personally I'm not completely sold on the matter for a variety of reasons myself, but the usual logic given for why you should be skeptical about it is terrible. The interesting questions are a great deal more complicated than something that can be dismissed with something that generally boils down to "Look, I just can't imagine the world changing that much, so it won't".
Re: Kurzweil's rebuttal to Paul Allen
#8The Singularity concept strikes me as a sort of wishful thinking. Technology advancing so fast that we no longer can control or understand it? Yeah, that already happened to my parent's generation with AOL, yet here I am texting this on my iPhone. New generations understand intuitively what the previous generation understood theoretically. Even so, I fully expect memristors to deliver strong AI.
> Even so, I fully expect memristors to deliver strong AI. Why would they? Is strong AI a function of storage capacity or speed? An AI running at 1/100th of what a future AI may be capable of is still an AI and I can't see how a mere improvement of a couple of orders of magnitude would do what decades of Moore's law have failed to do so far. If strong AI was just a matter of speed then we could theoretically take any…
Memristors are the silicon equivalent of neurons -- a time-dependent function with state. In mammals, the connectivity of millions of neurons enable the emergence of intelligence. I don't see any reason why a silicon-based neuron network of sufficient size isn't capable of the same.
Now, that's entirely different from purposeful design of an AI such that it speaks English and knows what I like for breakfast. I don't think we'll ever know how to sit down and write one in Notepad.
Instead, memristor-based AIs will be evolved using genetic algorithms or other evolutionary approaches. Yeah, maybe that's wishful thinking too, but that's how I see it playing out.
Re: Kurzweil's rebuttal to Paul Allen
#9Kurzweil makes some good points here. He doesn't address each and every criticism with his view of AI progress, but he does a good job calling out Paul Allen on not doing the homework.
Calling out someone for not doing their homework seems somewhere in the DH1-2 range: http://www.paulgraham.com/disagree.html There are some good points in Kurzweil's response, but the ones about Paul Allen are definitely not them. I think his best point was about extrapolating function from individual cells or structures, without needing to understand every single cell or structure individually.
Allen attacks Kurzweil rather directly in his essay and the essay title is even "The Singularity Isn't Near". So in that case I think it's fair game if Kurzweil calls him out for not even regarding the arguments he made in "The Singularity is near".
Re: Kurzweil's rebuttal to Paul Allen
#10Earlier quoted context omitted.
> Even so, I fully expect memristors to deliver strong AI. Why would they? Is strong AI a function of storage capacity or speed? An AI running at 1/100th of what a future AI may be capable of is still an AI and I can't see how a mere improvement of a couple of orders of magnitude would do what decades of Moore's law have failed to do so far. If strong AI was just a matter of speed then we could theoretically take any…
Memristors are the silicon equivalent of neurons -- a time-dependent function with state. In mammals, the connectivity of millions of neurons enable the emergence of intelligence. I don't see any reason why a silicon-based neuron network of sufficient size isn't capable of the same. Now, that's entirely different from purposeful design of an AI such that it speaks English and knows what I like for breakfast. I don't…
A memristor can be simulated by a couple of formulas and a few variables. (just like neural networks can be simulated).
In spite of our ability to simulate this sort of process for decades we have not succeeded in building a strong AI, the difference in having a hardware version is one of performance (just like Neural Network chips are typically a lot faster than their software simulated counterparts). There is no real difference in capability here, just a (possibly very large) speedup.
Now, I'm not ruling out that such a speedup will cause us to be able to create things that so far were not possible but I have a hard time convincing myself that this will almost certainly be the case.