[ExI] The symbol grounding problem in strong AI

BillK pharos at gmail.com
Sat Jan 9 21:45:48 UTC 2010


On 1/9/10, Damien Broderick wrote:
>  Isn't that exactly saying that you assign some kind of non-computable
> aspect to natural brains? (No reason why it should be strange, though.) As I
> said several days ago, a landslide doesn't seem to me to compute the
> trajectories of all its particles--at least not in any sense that I'm
> familiar with. We can *model* the process with various degrees of accuracy
> using equations, but it looks like a category mistake to suppose that the
> nuclear reactions in the sun are *calculating* what they're doing. I realize
> that Seth Lloyd and others disagree (or I think that's what he's saying in
> PROGRAMMING THE UNIVERSE--that the universe is *calculating itself*)
> but the whole idea of calculation seems to me to imply a compression or
> reduction of the mapping of some aspects of one large unwieldy system
> onto another extremely stripped-down toy system.
>
>  That might be wrong, I know. I hope Gordon knows it might be wrong as well.
>
>


I think what Gordon might be trying to say is that the brain is not a
*digital* computer.

Digital computers separate data and program.

The brain is more like an analogue computer. It is not like a digital
computer that runs a program stored in memory. The brain *is* the
program and *is* the computer. And it is a constantly changing
analogue computer as it grows new paths and links. There are no brain
programs that resemble computer programs stored in a coded format
since all the programming and all the data is built into neuronal
networks.

If you want to get really complicated, you can think of the brain as
multiple analogue computers running in parallel, processing different
functions, all growing and changing and passing signals between
themselves.

This modular parallel design is what causes 'consciousness' to be
generated as a sort of synthesis product of all the lower-level
modules. The digital computers we have today may not be able to do
this. We may need a new generation of a different kind of computer to
generate this 'consciousness'.

It is a different question whether we need this 'consciousness' in our
intelligent computers.


BillK



More information about the extropy-chat mailing list