On Jan 16, 4:28*am, Ian Parker wrote:

Yes, but what I say, technically at least, is backed by the majority
of academics. The dispute (amoungst academics that is) on the VN
machine is whether it is linked to nanotechnology or not.

The possibility of a von Neumann machine is not nutty. That self-
reproducing machines could go wild, macroscopic or microscopic, is
borne out by our experience with life, though. Saying that they're the
ONLY WAY to space exploration, though, is where you go off the rails.

Artificial intelligence does indeed have great promise in extending
our reach in space. The Spirit and Opportunity rovers, with limited
artificial intelligence, so that they can avoid following advance
orders from Earth into obstructions and pitfalls, illustrate this.

But chemical spaceflight is not so limited that it absolutely
precludes another route to space - establishing a self-reproducing
human colony in space, along the principles outlined by Gerard
O'Neill. That future, though, isn't absolutely certain either. Nuclear
power is known to be feasible, and requires vastly less initial
investment, than a space colony to crank out solar power satellites.

Both O'Neill and Zubrin, although having contributed a great deal to
the future possibility of extensive manned space exploration, were and
are also single-minded advocates of their respective visions for space
exploration. This is an obstacle to taking even *them* entirely
seriously.

Yes, Mars has everything that's needed in one place, even if it's a
bit low on nitrogen. But it's a planet, hence in a gravity well. The
lower technological demands of establishing a self-sustaining colony
on Mars need to be weighed against the lower benefits of one, compared
to getting started on O'Neill's vision.

And maybe the best way to get started on O'Neill's vision would be to
send teleoperated mining and manufacturing machines to the Moon,
instead of trying to put people there. The costs and benefits of
different alternatives will have to be carefully weighed.

The history of artificial intelligence shows us that it has been very
difficult to figure out how to program computers in more sophisticated
ways. So there's no guarantee of rapid progress in that field -
instead, slow progress has been the rule. Thus, even though progress
is slow in the mature technology of rocketry as well, arguing that
everything depends on AI is _prima facie_ unreasonable; if exploring
space is considered important, progress in it is not going to be
allowed to be dependent on one field in which progress is slow.

John Savard