3 ms·

The singularity won't happen: http://www.maximise.dk/blog/2008/11/why-singularity-may-never-arrive.html http://www.maximise.dk/blog/2008/11/why-singularity-may-

by mixmax 18y ago

The singularity won't happen: http://www.maximise.dk/blog/2008/11/why-singularity-may-never-arrive.html http://www.maximise.dk/blog/2008/11/why-singularity-may-neve...

Retric 18y ago
There are limits to computation, but nothing prevents computers from achieving human or super human level intelligence. I suspect that 50 years from now we will have human level AI for a wide range of things. We have already passed that point for Chess and one by one other things will fall. What really prevents the singularity is the limits of the physical world. The speed of light is already a problem such that PC ram is never going to drop in latency without moving it closer to the CPU. I don't care how smart you are you can't build a 101% efficient engine etc.
- listic 18y ago
  Are there any real signs of progress in computers that could lead to them achieving human level intelligence? Look, computer beat world champion in chess over 10 years ago, but is nowhere close to doing that in Go anytime soon: http://ru.ly/34 http://ru.ly/34
  - jcl 18y ago
    I wouldn't say "nowhere close", as the quality of computer go players has improved dramatically in the last year or so: http://games.slashdot.org/article.pl?sid=08/08/08/1243244&from=rss http://games.slashdot.org/article.pl?sid=08/08/08/1243244&#3...
  - Retric 18y ago
    Like CAPTA's? I would say the number of things computers can do better than people is growing. They are not unified in a single project but computers are gaining around 100x the computational ability every 10 years so the gap between 1% as smart as a human and human level intelegence can close rapidly. http://seedmagazine.com/news/2008/03/out_of_the_blue.php?page=6 http://seedmagazine.com/news/2008/03/out_of_the_blue.php?pag... Granted the human brain has around 100 billion neurons vs 10,000 neurons, with about 30 million synaptic connections between. However 100,000,000,000 / 10,000 is only 1/10,000,000 which is around 30 years unless the software becomes more efficient. PS: I don't think directly simulating a brain is the most efficient method of reaching human level AI but it's hard to see how it would fail.
- palish 18y ago
  "We have already passed that point for Chess..." No we haven't. Computers play chess by examining all possible outcomes, then choosing the optimal one. That seems to be how humans play chess too. But if you think about it, humans are much more spontaneous than that. We haven't succeeded in coding "spontaneity" yet.
  - dfranke 18y ago
    Spontaneity is just a synonym for "I don't understand what's going on when this happens".
rsheridan6 18y ago
There's no contradiction between Malthus and the coming of the singularity. Malthus essentially said that resources tend to be scarce, but there's no reason that they couldn't be scarce post-singularity. This line of thinking leads to doom scenarios, when you think about how scarce resources would be allocated (bottom line - probably not to humans), so it does contradict singularity-as-the-promised-land scenarios.
- mixmax 18y ago
  Scarcity means that the curve will break. The singularity is the notion that the growth will continue indefinitely at an ever wilder pace until it reaches a singularity. I'm saying it won't happen because technological advance will run into scarcity of resources, just like any other complex system. Malthus was just an example of this, since his thinking was much the same as Kurzweil's in regard to extrapolation of datasets. Whether this will happen before or after transhumanism I don't know. But like any other exponential curve it won't go on forever.
  - gjm11 18y ago
    "Singularity" here doesn't mean "infinite improvement", it means "such a big improvement that, from the point of view of people trying to predict it now, all bets are off". At least, that's how Vinge used the word and I think it's the commonest usage among present-day singularitarians like Yudkowsky.
    - pfedor 18y ago
      One thing I never got: sure, it's certainly possible that one day the technology will come into existence that will create beings (be it computers or genetically modified brains or whatever) that will be as superior to humans as we are to ants or rabbits. How on Earth does the singulatarian paradise follow from this premise? Doesn't it seem at least just as likely that those super-intelligent creations will eradicate humans, or enslave them, or do something to them which we don't have a concept for in our limited brains, and which is different from happy immortal existence?
Prrometheus 18y ago
The Malthusians have had egg of their face for the better part of two centuries now. Maybe one of these days they will be right. Still a pretty bad record overall, though.

by mixmax 18y ago

The singularity won't happen: http://www.maximise.dk/blog/2008/11/why-singularity-may-never-arrive.html http://www.maximise.dk/blog/2008/11/why-singularity-may-neve...

Retric 18y ago

There are limits to computation, but nothing prevents computers from achieving human or super human level intelligence. I suspect that 50 years from now we will have human level AI for a wide range of things. We have already passed that point for Chess and one by one other things will fall. What really prevents the singularity is the limits of the physical world. The speed of light is already a problem such that PC ram is never going to drop in latency without moving it closer to the CPU. I don't care how smart you are you can't build a 101% efficient engine etc.

listic 18y ago
Are there any real signs of progress in computers that could lead to them achieving human level intelligence? Look, computer beat world champion in chess over 10 years ago, but is nowhere close to doing that in Go anytime soon: http://ru.ly/34 http://ru.ly/34
- jcl 18y ago
  I wouldn't say "nowhere close", as the quality of computer go players has improved dramatically in the last year or so: http://games.slashdot.org/article.pl?sid=08/08/08/1243244&from=rss http://games.slashdot.org/article.pl?sid=08/08/08/1243244&#3...
- Retric 18y ago
  Like CAPTA's? I would say the number of things computers can do better than people is growing. They are not unified in a single project but computers are gaining around 100x the computational ability every 10 years so the gap between 1% as smart as a human and human level intelegence can close rapidly. http://seedmagazine.com/news/2008/03/out_of_the_blue.php?page=6 http://seedmagazine.com/news/2008/03/out_of_the_blue.php?pag... Granted the human brain has around 100 billion neurons vs 10,000 neurons, with about 30 million synaptic connections between. However 100,000,000,000 / 10,000 is only 1/10,000,000 which is around 30 years unless the software becomes more efficient. PS: I don't think directly simulating a brain is the most efficient method of reaching human level AI but it's hard to see how it would fail.
palish 18y ago
"We have already passed that point for Chess..." No we haven't. Computers play chess by examining all possible outcomes, then choosing the optimal one. That seems to be how humans play chess too. But if you think about it, humans are much more spontaneous than that. We haven't succeeded in coding "spontaneity" yet.
- dfranke 18y ago
  Spontaneity is just a synonym for "I don't understand what's going on when this happens".

rsheridan6 18y ago

There's no contradiction between Malthus and the coming of the singularity. Malthus essentially said that resources tend to be scarce, but there's no reason that they couldn't be scarce post-singularity. This line of thinking leads to doom scenarios, when you think about how scarce resources would be allocated (bottom line - probably not to humans), so it does contradict singularity-as-the-promised-land scenarios.

mixmax 18y ago
Scarcity means that the curve will break. The singularity is the notion that the growth will continue indefinitely at an ever wilder pace until it reaches a singularity. I'm saying it won't happen because technological advance will run into scarcity of resources, just like any other complex system. Malthus was just an example of this, since his thinking was much the same as Kurzweil's in regard to extrapolation of datasets. Whether this will happen before or after transhumanism I don't know. But like any other exponential curve it won't go on forever.
- gjm11 18y ago
  "Singularity" here doesn't mean "infinite improvement", it means "such a big improvement that, from the point of view of people trying to predict it now, all bets are off". At least, that's how Vinge used the word and I think it's the commonest usage among present-day singularitarians like Yudkowsky.
  - pfedor 18y ago
    One thing I never got: sure, it's certainly possible that one day the technology will come into existence that will create beings (be it computers or genetically modified brains or whatever) that will be as superior to humans as we are to ants or rabbits. How on Earth does the singulatarian paradise follow from this premise? Doesn't it seem at least just as likely that those super-intelligent creations will eradicate humans, or enslave them, or do something to them which we don't have a concept for in our limited brains, and which is different from happy immortal existence?

Prrometheus 18y ago

The Malthusians have had egg of their face for the better part of two centuries now. Maybe one of these days they will be right. Still a pretty bad record overall, though.