[ExI] What might be enough for a friendly AI?

[Keith Henson]

Re the whole subject, we have the ability to "look in the back of the
book" given that human exhibit intelligence.  (Sometimes I wonder.)  I
don't think the problem is as difficult at the hardware level as
people have been thinking.  I suspect that simulation at the cortical
column and its interconnections will be enough.  We also know that
brains are really redundant given that they degrade slowly as you keep
nicking chunks out of the cortex.  See William Calvin on this subject.

As far as the aspect of making AIs friendly, that may not be so hard
either.  Most people are friendly for reasons that are clear from our
evolution as social primates living in related groups.  Genes build
motivations into people that make most of them strive for high social
status, i.e., to be well regarded by their peers.  That seems to me to
be a decent meta goal for an AI.  Modest but with the goal of being
well thought of by those around it.

Eventually--if we can do even as well as nature did--a human level AI
should run on 20 watts.

Keith