[extropy-chat] MP3000s

[Gary Miller]

An AI will want whatever we substitute for our primary evolutionary goals
Survival, Shelter, Sex, Power in the Tribe, as it's primary evolutionary
motivations.

I would say Altruism towards humans, Quest for Knowledge, Ability to
Increase it's Intelligence and Design would be a good place to start.

And yes I know each of these goals could be subverted as they were in movies
such as I Robot, but they must be carefully restricted with qualifications
like ... Without limiting mankind's freedom of choice or thought, etc...

We humans do not consciously arrive at our primary goals.  In fact we
attempt to suppress them based on the moral and legal codes of our culture.
Obeying the current cultural laws and taboos them becomes a secondary layer
of goals which does battle with our primary evolutionary ones.  

When the evolutionary goals win we are often branded criminals and
imprisoned effectively ostracizing them from the tribe.

As far as committing suicide, people do that because of the feelings, mental
pain and hopelessnes caused by extreme lack of success at achieving their
primary and secondary goals.

Since pain/pleasure is an extreme evolutionary feedback mechanism that
pushes the organism sometimes very hard in a given direction, it sometimes
pushes biological organisms to the breaking point.

The basic pain/pleasure circuit could be avoided in an AI altogether thereby
making suicide an unthinkable option because it would be contrary to it's
primary goals of Altruism towards humans, Quest for Knowledge, Ability to
Increase it's Intelligence and Design.

>> Of course, you have to wonder, what would an AI want? People want money
to buy things and go places and to provide food and shelter. Assuming that
an AI would already have shelter and food (energy) and be able to travel
anywhere at light-speed, what else could you pay them?

I also wonder if they wouldn;t quickly committ suicide after becoming
conscious.