> From: Erik Reuter <[EMAIL PROTECTED]> > On Tue, Nov 19, 2002 at 04:25:19AM -0600, The Fool wrote: > > > From: Erik Reuter <[EMAIL PROTECTED]> > > > > > On Tue, Nov 19, 2002 at 12:17:35AM -0600, The Fool wrote: > > > > I doubt their calcutions for human brain processing power are > > correct. > > > > > > Their calculations are in the right ballpark; definitely they aren't > > > clearly incorrect. Here's a message Bob Chassell posted in 1999 that > > > gives a good background: > > > > A human mind isn't strictly sequential, like these kinds of computers > > are. Even using thousands of processors, they are really only solving > > sequential problems. A human mind is massively parallel processing. > > > > "10^11 neurons. Each neuron has about 5*10^3 synapses" > > > > Which comes out to about (5*10^14)! pathways. (this is an astronomical > > number, much larger than the projected # of subatomic particles in the > > universe). > > There is your mistake. It comes out to 1E11 * 5E3 = 5E14 pathways (or > possibly half that, I forget whether they count neurons at both ends of > the synapse or just one end)
But the number of possible pathways between any two nodes is the factorial. which is why it is very hard to optimise the shortest path a delivery person would take to 200 different deliveries. You have to calculate every single pathway, 200! > The article said: > > >will be able to complete 100 thousand billion calculations per second > >\227 a speed known as 100 teraflops that some scientists say is > >comparable to > > That is 1E14. It is in the right ballpark. Perhaps, but that was not what I was saying. > > Imagine for a second an internet where each node transmitted at the > > same speed, and each node was connected to not 1 other node, but 5 to > > 10 thousand nodes (like five to ten thousand individual fibre optic > > cables per node. > > No need to imagine. Just calculate correctly. _______________________________________________ http://www.mccmedia.com/mailman/listinfo/brin-l
