From: Philip Sutton (Philip.Sutton@green-innovations.asn.au)
Date: Sun May 30 2004 - 08:41:23 MDT
Why have more than one AGI?
Because each AGI type would be different at least to the extent that:
- their highest-order goal structures varied
- their thinking architecture varied
- their hardware base varied (especially in terms of grunt and speed,
inter AGI/human communications hardware)
Also each instance of an AGI would vary according to its training and
experience and what it is working on (major projects).
The fact that AGIs could have 'telepathy' and high bandwidth
communication, including of non-linear massively complex data sets
would not in my opinion change the significance of the above because
even an AGI is going to have to do work to integrate info received from
another AGI and when the info integration has occurred the AGIs are
still not identical. In my opinion mind-melds would not obliterate the
unique characteristics of AGIs.
So why have more than one AGI? Because they would have different
strengths and weaknesses that would make them better suited to
different tasks and because they could provide each other with second
opinions/alternate perspectives/reality checks and could work
collectively to manage risks arising from aberrant AGIs.
I've thought about whether I'm anthropomophising or biology-mophising
and I can't see any reason to think that I am.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:47 MDT