[ExI] Brain emulation, regions and AGI [WAS Re: Kelly's future]

[Kelly Anderson]

On Fri, May 27, 2011 at 7:18 AM, Richard Loosemore <rpwl at lightlink.com> wrote:
> There is actually no true modularity, there are just regions with apparent
> specializations, which are more or less definable or distinct, depending on
> the case.  The terms used are also not uniform:  many are just "areas" (as
> in Brodmann areas), but that implies cortex, whereas some significant chunks
> or sub-cortical.

Thanks. That helps me to understand some things better. My
understanding is that there are some divisions that are pretty
structural, such as the brain stem... but if I understand what you're
saying it is that in the neocortex, there aren't such sub organs.
Right?

> My own, more general answer to the issue of AGI via brain emulation is that,
> as you are suggesting, the modeling of the human mind/brain is likely to be
> the first successful AGI, but this is unlikely to be whole brain emulation
> in the sense of a low-level neuron-by-neuron copying.

I agree 100%. In implementing this whole brain emulation, there may be
areas where the best initial emulation is neuron by neuron emulation,
but I hope that's just a small part of the emulation. And neuron by
neuron emulation seems like a valuable first step in figuring out what
the higher level functions are and how they work, I think.

-Kelly