4 ms·

Isn't a lot of the singularity idea based on flawed math? I always hear about and see graphs of exponential curves about to "go vertical", but... exponential c

by simon_ 17y ago

Isn't a lot of the singularity idea based on flawed math? I always hear about and see graphs of exponential curves about to "go vertical", but... exponential curves definitely don't do that. You can zoom in on any point of e^x and it will look like it's about to have a singularity... In more general terms... if ~technology has been getting better with some short doubling time for hundreds of years... why is it that the NEXT doubling is supposed to be the really significant one?

ivankirigin 17y ago
There is a big difference between humans iterating to improve technology and machines iterating to improve technology that improves the machines. The latter is a stronger feedback loop. But generally, I agree. We're biased to under-appreciate the impacts of past changes.
rms 17y ago
The last doubling was really significant. You know, instant free communication to humans anywhere on the world as well as instant access to the sum total of human knowledge.
- simon_ 17y ago
  Right, and the next doubling will be even better (twice as good!). But the sense that we're right on the threshold of the infinite future has probably been around at least since industrialization. For example, that bit about instant access to people and knowledge was probably said about the telegraph too.
  - rms 17y ago
    Sure, I agree. My only point is that people seem to underestimate the regular, non-singularity exponential doubling. Already some aspects of science and civilization have become more far-fetched than the sci-fi of 50 years ago. Like http://www.wired.com/magazine/2009/10/mf_optigenetics/ http://www.wired.com/magazine/2009/10/mf_optigenetics/
- tel 17y ago
  I bet that whole light coming from little threads in your walls that kill you if you touch them was quite a doozy too. Commercial airflight. Fertilizer. The idea of foreign pathogens causing diseases. Which is not to say that internet technology isn't widely more crazy than all of those previous or that there won't continue to be wider and crazier things to come. It is to say that we're not escaping exponential growth and there still isn't anything like a spike in that sort of curve. The singularity theory proposes an increase from exponential: some technology that speeds new technology's ability to speed new technology. Without the feedback --- without the efficient feedback --- everything "stays exponential". Which is only really depressing if you're looking for hypersmart AI and endless free human life in the next 20 orso.

by simon_ 17y ago

ivankirigin 17y ago

There is a big difference between humans iterating to improve technology and machines iterating to improve technology that improves the machines. The latter is a stronger feedback loop. But generally, I agree. We're biased to under-appreciate the impacts of past changes.

rms 17y ago

The last doubling was really significant. You know, instant free communication to humans anywhere on the world as well as instant access to the sum total of human knowledge.

simon_ 17y ago
Right, and the next doubling will be even better (twice as good!). But the sense that we're right on the threshold of the infinite future has probably been around at least since industrialization. For example, that bit about instant access to people and knowledge was probably said about the telegraph too.
- rms 17y ago
  Sure, I agree. My only point is that people seem to underestimate the regular, non-singularity exponential doubling. Already some aspects of science and civilization have become more far-fetched than the sci-fi of 50 years ago. Like http://www.wired.com/magazine/2009/10/mf_optigenetics/ http://www.wired.com/magazine/2009/10/mf_optigenetics/
tel 17y ago
I bet that whole light coming from little threads in your walls that kill you if you touch them was quite a doozy too. Commercial airflight. Fertilizer. The idea of foreign pathogens causing diseases. Which is not to say that internet technology isn't widely more crazy than all of those previous or that there won't continue to be wider and crazier things to come. It is to say that we're not escaping exponential growth and there still isn't anything like a spike in that sort of curve. The singularity theory proposes an increase from exponential: some technology that speeds new technology's ability to speed new technology. Without the feedback --- without the efficient feedback --- everything "stays exponential". Which is only really depressing if you're looking for hypersmart AI and endless free human life in the next 20 orso.