Uploading terms

Algimantas Malickas (malickas@pub.osf.lt)
Mon, 31 Mar 1997 13:42:47 +0300 (LDT)

I will try evaluate the possible terms of morphological and gradual

In 1946 one MIPS cost about 150 M$, in 1981 - 10 k$, in 1995 - 10 $. After
simple calculations I have got, that mean annual coefficient of computational
power growth was:
1.32 times/year - during 1946 - 1981 years period,
1.64 times/year - during 1981 - 1995 years priod.

I believe, we can affirm that brain power is between 10^13 and 10^18 sinapses
per second. Let us assume, that 1 sinapse processing require 10-1000 floating
operations. So, the necessary computation power will be 10^14 - 10^21 FLOPS.

If we would take 1981 - 1995 coefficient (1.64) and most powerful current
computer (2*10^12 FLOPS) we will got that brain power computer will arose
between 2005 and 2039 years.

The brain power computer is not sufficient condition for morphological uploading,
of course. Scaning, decoding and ethical problems can very postpone these
possibilities. But for humanoid AI (according to Carl Feynman classification)
these problems will not so bigs. In this case, we can expect, that human
power AI would arose earlier than uploading possibilities. I have some doubts
about slow evolution human and fast evolution AI co-existence possibility.
I think that during 10 - 20 years such co-existence the human society would
have a big danger of degradation.

So, I think, the full human uploading should be possible during 2005 - 2060
years. In other case we have a chance, that it will not be possible never.

More optimistic I see the gradual uploading perspectives. We have a good chance,
that sufficiently brain - computer interface will be created during 0 - 10
years (if even one current project will be successful). The necessary computer
power would be lessen than in the morphological uploading case. In any case,
I think, the gradual uploading will not be implemented later that humanoid AI.
It can be implemented almost at once after necessary computers creation (I
believe about 2015).

Still more, already now we can expect some seriously results concerning
gradual uploading to the other brains. Let the brain-brain interface is
incomplite, let these experiments would be with animals only, some urgent
results is possible in the nearest future.

Best regards,