RE: Seed AI milestones - Explicit Knowlege Codification

From: Ben Goertzel (ben@goertzel.org)
Date: Wed Feb 27 2002 - 15:15:43 MST


> On Wed, 27 Feb 2002, Samantha Atkins wrote:
>
> > More importantly, it has an efficient "learning machine" built in. The
> > simpler this "learning machine" is, the more likely we humans will be
> > able to build something at least as good.
>
> Human baby as a learning machine is an excellent learning machine, but
> it's a very complex one. Expectations to build a simple gadget that just
> magically extracts knowledge from arbitrary sources is unreasonable.
>
> The simplest approach to that would be to copy a human baby in machina.

If

1) one had the right type of hardware to create a human-baby-like body as
well as brain

2) one had a detailed map of the human brain and body to use as a blueprint
for constructing a particular baby-like system using this type of hardware

then YES, you're right, that approach would be simplest.

However, neither of these conditions holds. Thus it is not at all clear to
me that the approach you propose is simplest.

Just because there is no adequate general-intelligence design in the
published research literature, or in your own mind, does not prove such a
thing does not exist ;->

-- Ben g



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:37 MDT