From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Sat Jun 22 2002 - 03:11:02 MDT
Eugen Leitl wrote:
> On Sat, 22 Jun 2002, Mike & Donna Deering wrote:
>
>
>>Estimates of how much conventional computing hardware is required for
>>human level intelligence based on observations of the human brain have
>>been seriously inflated for the following reasons.
>
>
> Of course you realize that the trend in growth of past estimates done by
> people with otherwise impeccable accomplishment track speaks against your
> statement. Granted, though smart people, they had no idea of modern
> neuroscience. We might be not as smart, but we have several decades worth
> of literature more to study.
>
> If you track the neuroscience publications more or less closely you'll
> find little basis for optimism there. It's not just the complexity which
> makes accurate numerical models of wet systems a stark nightmare, it's the
> number of bells and whistles, and how much of systems intrinsic artifacts
> are being hijacked for implementing features.
>
> The trend does not look good here.
I find myself agreeing with Eugene on this score. Many of the computational
models on which estimates of the "maximum hardware power" of the human brain
are based seem absurdly simple by comparison with any modern paper on
computing in actual human neurons. At the same time, the notion that all
that potential power is being used effectively - that all systems intrinsic
artifacts contribute materially to computation at higher levels of
organization, such that it would require a perfect simulation of that
systems artifact to obtain that contribution through any other means - is
equally absurd. I would place the estimate of computing power needed for a
blindly faithful upload as being at least on the order of 10^19 ops/sec or
10^21 ops/sec. For AI, on the other hand, the number of major differences
at all levels of organization is enough to shatter the analogy. I'm not
sure that there exists anything to be gained from the mapping any more. But
I do agree with Eugene that existing estimates of supposed "maximums" are
not taking into account anything like the complexity of computing in single
neurons.
>>Therefore, I estimate, that the $1000 desk top of 2002 is not more
>>than three doublings away from human level complex problem solving
>>general deliberative software implementation capability.
>
>
> You're still confusing crunch (measured in what? remember memory
> bandwidth) with a generally intelligent system. You'll observe that a kg
> of neuron suspension in liquid culture is usually not very intelligent.
>
> Even if we assume you're not wildly overoptimistic, you're completely
> ignoring the bootstrap issue. It may require a lot of crunch to be able to
> do neat stuff with a modest amount of crunch.
I don't see how Mike Deering gets from his statements to the above estimate.
It seems like a non sequitur.
-- Eliezer S. Yudkowsky http://intelligence.org/ Research Fellow, Singularity Institute for Artificial Intelligence
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:39 MDT