RE: AGI funding (was Re: Some bad news)

From: Ben Goertzel (ben@goertzel.org)
Date: Fri Nov 08 2002 - 10:10:06 MST


Rafal wrote:
> ### How much of a difference, conceptually, is there between Biomind and
> AGI? The structure of data in gene banks is IMO remarkably
> similar in terms
> of its messiness and multidimensionality (expression, structure, sequence,
> refs to research on function and clinical outcomes) to the kind of data an
> AGI would need to analyze while trying to understand the world. I realize
> that many features that are analogous to human personality (rather than
> cognition), the goal system, would be absent from a narrow AI, but still I
> feel you should be able to get a lot of AGI mileage from Biomind. Am I
> right?

Yes, we believe we can get a lot of AGI mileage from Biomind.

Many of the cognitive processes are the same for both, and analyzing bio
data has a lot of similarities to any other form of "perception"

What's missing is:

-- action, in any serious sense [until we have Biomind controlling the
robots doing bio experiments, which is not a totally fanciful notion, but is
not in our near-term game plan]

-- coordination of perception, cognition and action

These are rather large omissions, obviously!

To make the comparison a different way: Of the 25 or so basic
"mind-processes" in Novamente, Biomind's initial software will use about 5,
and could ultimately use about 15.

Biomind can help us create, test and tune a lot of the components required
for AGI -- but that's still not AGI ...

-- Ben



This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:41 MDT