[ExI] Limiting factors of intelligence explosion speeds.

[John Clark]

On Jan 24, 2011, at 1:27 PM, Richard Loosemore wrote:

> I don't give a fig if some cellular automaton might do in the next 10 gigayears

Wow, 10 billion years is a long time, but let's see, the signals in the human brain move at about 10 meters a second, in a AI they would move at 300,000,000 meters a second, 30 million times faster, so it would take the AI a little over 300 years to do it, that's still a long time. But the AI would have more hardware not just faster hardware working on the problem. The human brain is about 15 cm across so a AI could have a brain 45,000 meters across with no more delay between one part of its brain and another than we have with our human brain. With Nanotechnology you should be able to fit the complexity of a human brain into one cubic centimeter. So in a sphere with a radius of 22,500 meters you could fit in 2.7 *10^19 of them. So it will take the AI .35 nanoseconds to solve your 10 billion year long problem. Of course with cooling and other engineering problems you might only be able to pack a tenth as much stuff in, then it would take a glacial 3.5 nanoseconds to bring on the singularity.

  John K Clark

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20110126/27af6022/attachment.html>