[ExI] AI Motivation revisited.

[Kelly Anderson]

2011/7/1 john clark <jonkc at bellsouth.net>

> On *Thu, 6/30/11, Stefano Vaj <stefano.vaj at gmail.com>* wrote:
>
>
> "What I think is reasonable, and corresponds to everyday experience, is
> to expect that a PC in comparison with other systems, such as organic
> brains, be incredibly faster at some tasks (say, arithmetics), and
> incredibly slower at some other (say, pattern recognition or neural
> network emulation)."
>
> Organic brains may work by neural networks but they are no better at
> emulating one than a non organic metallic computer, in fact they're not
> nearly as good; if you doubt that then look at a very very small network of
> only a few hundred elements and through pure visualization (or "emulation"
> if you prefer that word) predict how this network will behave. It's hard as
> hell, we're just not very good at that; electronic computers can do it much
> better.
>
> As for pattern recognition we still have an edge over the machines, but I
> doubt our superiority will last another decade.
>
>
One way to look at this is that when we "emulate" computers, by say doing
math in our head, we do it at a rate that is thousands (if not millions) of
times slower. Similarly when computers emulate the pattern recognition tasks
that we are good at, they are thousands of times slower than us...

The difference of course is that computers keep getting faster, and we stay
the same (so far). Therefore, even if computers are thousands of times
slower at pattern recognition, someday computers will be thousands of times
faster than they are today, and then eventually a trillion times faster.

The outcome of this race is inevitable.

-Kelly
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20110701/7aeeb78c/attachment.html>