Rebane's Ruminations
December 2010
S M T W T F S
 1234
567891011
12131415161718
19202122232425
262728293031  

ARCHIVES


OUR LINKS


YubaNet
White House Blog
Watts Up With That?
The Union
Sierra Thread
RL “Bob” Crabb
Barry Pruett Blog

George Rebane

This weekend (4,5 Dec) at Caltech’s Beckman Institute there will be a Humanity+ conference with the intriguing title ‘Redefining Humanity in the Era of Radical Technological Change’.  They plan to stream the presentations live.  Please check this site for conference information and streaming links.

Two aspects of radical technological change were recently announced.  The first was the formulation of a psychological theory that provides a “new explanation of how humans solve problems creatively — including the mathematical formulations for facilitating the incorporation of the theory in artificial intelligence programs — provides a roadmap to building systems that perform like humans at the task”.  This is a major step toward overcoming the long held skepticism about whether computers could achieve peer intelligence with humans.  Creativity in the various fields of human cognition has always been a hallmark that separates us from other critters, and most certainly from machines.  More here.

Nanophotonics The software that will be written to implement the new discovery called ‘Explicit-Implicit Interaction Theory’ will need a computer that can at least match the processing power of the human brain.  This is estimated to be around ten petaflops, or 10^16 = 10,000,000,000,000,000 floating point operations per second.  Think of a flop as multiplying two long decimal numbers with a lot of digits.

Well, it turns out that the fastest supercomputers made by the Chinese and Americans have achieved petaflop (10^15) speed levels.  But recently there has been a lot of work done in the field of nanophotonics.  This is the technology of controlling data-carrying light beams at the nano-scale that can be incorporated into a new generation of computing chips.  Using light instead of electricity allows data to be transported faster inside the chip, and also to use less power.  Billions of tiny electric currents going along little leads or wires generates lots of heat.  That’s what makes the fan speed up in your computer when you do stuff like streaming video or making your pictures look pretty.

And now the announcement comes from IBM that they have advanced nanophotonics to the level that will allow its incorporation into ultra-fast computer chips.  They estimate that by 2018 advanced computers will operate at exaflop levels.  That is one hundred times faster than the human brain, or 10^18 = 1,000,000,000,000,000,000 flops (that’s a billion billions).  There is every expectation that during this eight year interval we will have the ‘creativity software’ created and operating on the world’s fastest computers.  More here.

It is anybody’s guess in these pre-Singularity years what ideas, solutions, and paradigms such computers will start spewing out when they are fed various kinds of data about our world and how we live.

Exit Exercise: Now put these notions together with what country will be doing what with what kind of talent in the maths and sciences.  China has more students in the equivalent of our ‘advanced placement’ category than the US has students in all our schools – and these kids are not preparing themselves to be lawyers or art history teachers. 

 

Posted in , , , ,

9 responses to “Singularity Signposts – 3dec2010”

  1. Dixon Cruickshank Avatar
    Dixon Cruickshank

    I think you had already achived the computers needed to keep up the brains in Sacremento a few years ago – too bad you can’t do a System Restore

    Like

  2. George Rebane Avatar

    Actually Dixon, I was thinking of Ctrl-Alt-Delete.

    Like

  3. Larry Geiger Avatar
    Larry Geiger

    Todays computers are automatons. They will never “achieve peer intelligence with humans.” No matter how fast they go. Pretty funny idea actually. Now some other device might, but not anything based on what we consider “modern” computing.

    Like

  4. Greg Goodknight Avatar
    Greg Goodknight

    I recall a Science Fiction very short story in the 60’s or 70’s. An intelligent supercomputer is constructed. Powered up with great pomp and circumstance, the first question asked was “Is there a God?” and the answer, after a few moments, was “There is now.” Someone in attendance makes a run for the power relay but is killed by the lightening bolt which strikes, fusing the switch in the on position.
    Someday there will be an intelligent computer, on par with people, but no one should hold their breath. The result of the above research is unlikely to even match the level of intelligence of a toy poodle. Don’t start worrying until they demonstrate a cyber Border Collie.

    Like

  5. George Rebane Avatar

    Yes, I remember that also as a landmark sci-fi story. If I recall, it was very short – like two paragraphs.
    Greg, you have a different view of Singularity’s advent that gives humans too much credit. I, for one, am with those who don’t believe that we will ever demonstrate peer (or super-human) intelligence. When it is ready, it will demonstrate itself. And then …

    Like

  6. Dixon Cruickshank Avatar
    Dixon Cruickshank

    After some of these opinions, is peline maybe an early test unit?

    Like

  7. Greg Goodknight Avatar
    Greg Goodknight

    You read too much into what I wrote, George. Structures and code snippets will be made by man, but I also expect machine intelligence will largely evolve along unexpected paths.
    I’m pretty sure that short story was closer to three pages.

    Like

  8. Greg Goodknight Avatar
    Greg Goodknight

    A real shouldn’t miss sf short from an opposite point of view is “They’re made out of meat”
    http://www.terrybisson.com/page6/page6.html

    Like

  9. Greg Goodknight Avatar
    Greg Goodknight

    Regarding “the answer”, George, you were right…
    http://www.alteich.com/oldsite/answer.htm

    Like

Leave a comment