[extropy-chat] Timescale to Singularity

Marc Geddes marc_geddes at yahoo.co.nz
Fri Jun 17 06:16:18 UTC 2005


>A noble sentiment, but I'm not sure 2030 is realistic; my guess, for
>what it's worth, is that mid to late 21st century is more plausible
>than early. Only time will tell, of course, but remember it's a
>marathon, not a sprint: if 2030 comes around and the world's problems
>are still with us, that does not necessarily mean we have failed to
>solve them - only that we have not done so _yet_.
>
>- Russell
 
For a variety of reasons I think 2030 is a very conservative, realistic and achieveable target for Singularity.
 
I base my target date on comments made by leading AI researchers such as Wilson, Yudkowsky and Goertzel (who all stated on various messageboards that the problem of general intelligence was now mostly solved and it is only the Friendliness problem that now awaits solution.  Goertzel was also sure that his Novamente project could be coded in only 30 000 lines of a high level language, which is only a modest size).  Aside from that there is all the evidence for accelerating progress in info-tech, summarized well in Damian Broderick's 'The Spike', Moore's Law etc.  There is also concrete evidence to support this rate of progress in the form of the ferocious rate at which news is coming out of the labs (for instance millionare Jeff Hawkins project, or the recent 'Blue Brain' project to simulate the brain to name just two recent impressive examples).   
 
In any event, there is also evidence that other technologies such as nano-technology or bio-technology should be reaching maturation by 2030 and the dangers posed by these technologies require AGI to handle saftely.  
 
Finally, for me personally, I'm 33 now and by 2030 I'd be 58, which is too old to be sure of my further survival without major advances in bio-tech (survival rates plummet after one turns 60).
 
So it all looks like 2030 really is the end of the line (at least as far as I'm concerned).
 
All that said, I concede that trying to predict target dates for things is fun but can all too easily degenerate into mental masturabation and pointless philosophisizing).  Rather than prediction, one is really *projecting* or *goal setting*
 
I do agree with you that we're probably not nearly as close as the hard-core enthusiasts think.  Real AGI for instance, probably requires major advances in probability theory and a mathematics of complex systems which doesn't exist yet.   For instance a while back I stumbled upon a new kind of epistemology called 'Procedural Naturalism' which goes beyond Bayes and I also realized that deductive reasoning has never been properly incorporated into probabilistic causal networks.  So there are hints there that new maths may be required.   All ball-busting stuff.
 
Finally there is the big problem of morality yet to be solved.     It's one thing for people like Wilson and Yudkowsky to spout on and on about how clever they are to know the solution to general intelligence, but they got most of the answer by reading it in numerous papers and books.  Easy to absorb what others have done.  Far far harder to forge into new terrority.  There are no books and papers for them to copy the solution to morality from, so I predict that they're up the creek without a paddle when it comes to morality ;)   Serves them right really.   Monumental arrogance displayed to others, over-confidence, too much egoism etc.
 
 
 
 
 
 
 
  
 
 
 
    


---

THE BRAIN is wider than the sky,  
  For, put them side by side,  
The one the other will include  
  With ease, and you beside. 

-Emily Dickinson

'The brain is wider than the sky'
http://www.bartleby.com/113/1126.html

---

Please visit my web-site:

Mathematics, Mind and Matter
http://www.riemannai.org/

---
Send instant messages to your online friends http://au.messenger.yahoo.com 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20050617/e9df295f/attachment.html>


More information about the extropy-chat mailing list