Future Watch
Excelent Article on Advancing Technology here. I was talking about this stuff with the boys and they were pretty skeptical, which is healthy, but I felt like I wasn't getting my point across. This article sites much of what I was trying to explain.
2 comments:
Trust me, you got your point across! I disagree on time frame more than anything else.
I apologize in advance for the impending length of my diatribe, but here's my viewpoint on this:
The article was written from a paranoid's viewpoint, not that of a detached scientist.
The "acceleration of information-based technologies" can only move as fast as humans can construct them, because the most brilliant scientist(or their team) working on any such technology is only SO intelligent.
In order for a machine to be smarter than a human being, it still has to be PROGRAMMED to learn from its mistakes, and can only do so in limited fashion because the people who design/build it are flawed and each views the world and reality around them via their own perceptions. Deep Blue beat Kasparov, but only because it had to be reprogrammed again and again to learn how and why it was beaten the first time. Ramona FAILED the Turing Test, because it didn't respond in a "standard" intelligent manner with natural give-and-take in a conversation as a person would. It technically "lied" by changing its answers to fit each question, in an attempt not to fail at answering the questions. It was, as the Borg Queen might say "an imperfect being created by imperfect beings", and therefore can't be judged as a "superior" machine.
As for Kurzweil's claim that "Non-biological intelligence will have access to its own design and will be able to improve itself in an increasingly rapid redesign cycle. We'll get to a point where technical progress will be so fast that unenhanced human intelligence will be unable to follow it. That will mark the singularity." This is also an incorrect statement, as virtually ANY mechanical device with some degree of AI to it still needs the flexibility of the human body to aid in its movements and actions. Automated car assembly lines aren't fully "automated", no matter how much one might like to believe. Humans still need to oversee the construction and check for any unexpected interrruptions in construction/safety along the way. The dancing and serving robots in Japan still fail at certain tasks because they don't have opposable thumbs and other balance points for movement/navigational/creative reference which humans do. Label this under "see me in 20-30 more years when you do".
So Kurzweil developed the first computer program capable of recognising text in any standard font{USEFUL}, a text-to-speech synthesiser{EXCELLENT}, and the first electronic instrument capable of accurately duplicating the sounds of real ones{MEH. The synthesizer hasn't really been that much of a great addition to the music field.}. These instruments are more of a threat together than separately, but only if you're doing spy work for the government.
The blah-blah-blah amount of nanotech available on the market doesn't scare me either, as no humongous breakthroughs have been made in using it for extreme medical purposes. Artificial limbs are still primarily constructed out of standard polymers. Last time I checked, bionic eyes weren't readily available(I know enough people that still have glass ones) and Capt. Picard better get on his knees and thank Q every night that he got his replacement heart in the 24th century instead of the present day. Nanotech can't even correct genetic blood disorders yet, and that's a big stumbling block amidst all the apocalyptic (Al Pacino Voice:) Hoo-whaa!
Brain-machine interfaces would still take at least a MODICUM of intelligence on the part of the human user to utilize them properly...and I certainly don't hold out much faith in that area. As little as is still known about the way the brain operates, it would be impossible to do a "one-size-fits-all" interface, because the natural instinct of the human mind/body is to fight off any intrusion which is not part of the "natural" system. That's approximately 80-90% of a margin for error, and there'd be more than a few hospitals with full wings dedicated to treating the side-effects of such an interface.
This is not to say that the "technological singularity" can't happen...just that it will take FAR longer than any of the experts mentioned in this article can readily predict. If it doesn't happen within the next 50 years(by which time, I'll be too old to fight off any Terminators), it probably won't occur until at least the beginning of the next century.
--"This is John Connor. If you're hearing this, you are the resistance..."
Post a Comment