Hugo de Garis is interviewed by Wired and describes his views of the AI
future. He foresees the creation of super-intelligent AIs, "artilects"
(artificial intellects). The race to create these AIs will be, along
with the nano race, a principle form of competition in the next century.
Some will favor creating these machines, which are arguably our
replacements: the Cosmists. Others will oppose them, the Terrans.
de Garis sees the future somewhat similar to Moravec, with the Cosmists
escaping from the Earth, dominated by Terrans (rather confusing choice of
terminology there).
Moravec suggested that this would lead to temporary stability, with
the AIs in space leaving the earth alone for some time while their own
skills matured, but de Garis sees a more violent short-term outcome.
The Cosmists will seek refuge in space where they can build their
artilects, while the Terrans will pursue them with intent to destroy.
The Cosmists will need to develop their own offensive capabilities as
quickly as possible in order to counter the threat from the Terrans.
Meanwhile de Garis is working on building a million neuron milli-brain
based on a new configurable cellular automaton chip from Xilinx. Then
he'll use evolution (or should we call that directed evolution?) to try
to make the brain smart.
Sounds like he's got a ways to go before his machines take over.
Hal