Claes Frisk, Claes.Frisk at psyk.ks.se, wrote on Mon, 30 Aug 1999:
> Are there any reasonable estimations of the capacity
> (speed, memory,...) of the brain in computer terms?
Three major vectors of comparison can be eigen-Mashey-ed as:
1) speed in logic-gatings per second;
2) memory in:
A) absolute size;
B) life-span accommodation.
3) degree of parallelization.
> I realize that the two can't be directly compared,
Sez who? They are already on a collision course, according to
> but I guess that it's been done a number of times anyway.
Once more into the breach, the "cerebrum hominis sapientis"
crawls along at a neuronal propagation speed of what? 200 m/s?
And the logic gatings, as our hero J.C. Eccles (1903-1997)
discovered, must slow down and switch from electrical to
chemical transmission across the synapses where nerves "touch"
(which is the root meaning of Greek synapse: touching together).
So the human brain has lost the comparison like an election in
its early returns (say, that of President George "Cocaine" Bush
versus President Albert "Wooden" Gore -- a child's role model!?),
but like a Phoenix arising from dust storms and ashes the CNS
goes into comparative warp speed by means of 100 billion neurons
(versus computers with CPUs of 10-100 M transistors and memories
of either the 32-squared limit or nowhere near the 64-squared lim.)
http://www.tera.com Tera Computer with its taped-out MultiThreaded
Architecture (MTA) on a 64-bit chip supports up to 128 instruction
streams (or virtual processors) versus Massively Parallel Process-
ing (MPP) already up and running in your newsgroup reading wetware.
is an all-out no-nonsense no-cocaine effort to recreate the brain
"in silico" with a wooden adherence to such brain highway building
traditions as simulating the entire cortical sheet with a mindgrid
of arrays (psi, uk, ear) to contain concepts, English and phonemes.
The recent Tokyo traffic jams resulting from the first "roll-over"
of timekeeping in GPS history serve notice on the computer industry
to please provide for us public-domain mindmakers an architecture
independent of actual starting-points in history (are babies stamped?)
so that the legions of mind-coders may vie for the distinction, nay
the *necessity* of running an AI from a watch-out-it's-coming point
in time and never again till the end of time snuffing out the alife.
> [...] Claes Frisk
> Department of Clinical Neuroscience
> Karolinska Institutet
The medallion-granting outfit -- for all you would-be Nobelists.
> +46 8 517 71728
>Claes.Frisk at psyk.ks.se [...]