Re: (de Garis') are fake neurons

From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Wed Jan 26 2000 - 17:42:18 MST


Eugene Leitl wrote:
>
> The tasks of creating an AI and
> neuronal emulation are only distantly related. Neuronal emulation is
> distinctly harder.

Hear, hear! In fact, I would divide this into three different tasks:

1) Getting useful AI out of a neural net;
2) A neuronal emulation good enough that it's helpful to use synaptic
patterns and neuroanatomical architectures and training methods derived
from study of the human brain;
3) Emulating an actual human, or useful parts thereof, that you somehow scanned.

-- 
               sentience@pobox.com      Eliezer S. Yudkowsky
                  http://pobox.com/~sentience/beyond.html
Typing in Dvorak         Programming with Patterns  Writing in Gender-neutral
Voting for Libertarians  Heading for Singularity    There Is A Better Way



This archive was generated by hypermail 2b29 : Thu Jul 27 2000 - 14:02:45 MDT