From: Samantha Atkins (samantha@objectent.com)
Date: Thu May 27 2004 - 14:42:08 MDT
On May 26, 2004, at 5:40 AM, Ben Goertzel wrote:
> Samantha,
>
> I agree that the ethics of enhanced humans will be easier to predict
> than the ethics of self-improving AGI's.
>
> However, the ethics of AGI's produced by enhanced humans will not
> necessarily be easier to predict than that of those produced by
> ordinary
> humans. That depends on the orientation of the enhanced humans: are
> they trying to make more predictable AGI's or are they serving some
> other purpose in their AGI design?
>
Degree of intelligence counts. If the humans building/training the
AGIs are brighter then they are more likely to succeed and more likely
to succeed in instilling Friendliness if that is their goal. As I
know you have build quite sophisticated systems I know you are
painfully familiar with the limits of what even very bright people can
accomplish in a given amount of time. You are quite aware of where
the limits are.
> What this thread really leads to is the (obvious) conclusion that we'd
> be better off to enhance humans for IMPROVED WISDOM before we enhance
> them for improved intelligence.
>
We need both. One way to proceed is to find relatively wise people
and enhance them (if it seems wise to them).
> However, this seems fairly unlikely to happen, because wisdom, rather
> than intelligence, will be seen as providing less competitive advantage
> to the parties pushing for genetic modification of humans.
>
I am not speaking of embryonic genetic enhancements. There are some
gene therapy regiments beginning to appear that seem (in lab mice at
least) to enhance memory by roughly 30% and cut learning time in half.
If such works for humans then it would be a candidate. There are
also some new drugs just coming to market that have significant
beneficial effects. Beyond this I believe a lot can be done with
computer-based (wearable soon) of human information handling abilities
and boosting memory, association, integration, logic computation and
grasping of larger patterns.
On the wisdom side of it I am less sure what can be done. I think many
can be made wiser by even very old techniques. But it takes time and
commitment that few are ready for. Plus the older techniques are
buried often beneath a lot of superstition and mumbo-jumbo. That
turns many off. I don't have a lot of hope for many to become wiser
quickly. I do have some hope that a more positive and helpful set of
memes can capture the minds and hearts of many.
> "Genetically engineering a wiser human" is an excellent transhumanist
> project. If I weren't so busy with AGI maybe I'd take it up myself.
> Hopefully some of the things I'm doing with bioinformatics at Biomind
> will ultimately be useful for this goal ... in the unfortunate
> circumstance that we fail to create a benevolent AGI before biotech
> advances to the point enabling the creation of wiser OR cleverer humans
> by genetic manipulation, that is.
>
>
We might create highly enhanced humans by accident in reaching for AGIs
as we are likely to want pretty tight HCI to the young AGI in its
training.
- samantha
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:47 MDT