Re: Geoffrey Miller on Fermi

From: Russell Wallace (
Date: Wed May 03 2006 - 09:55:33 MDT

On 5/3/06, Dani Eder <> wrote:
> Once you have automated manufacturing systems capable
> of making copies of themselves, all it would take is
> one person deciding to do it. By automated
> manufacturing
> systems, I'm not necessarily talking about nanotech
> assemblers. I'm talking about automated machine tools
> and robots.

Maybe so. The problem is getting to that point.

The DARPA Grand Challenge cross-country robot races
> show we are very close to having autonomous mining
> robots.

Unfortunately no. Us computer and space people are prone to thinking that,
because we don't want to acknowledge the complexities involved; to get a
more realistic picture, you need to talk to mining and manufacturing people.
Here's an example of the feedback when someone (Monte Davis, on did:

I don't know if that link will work correctly, so I'll copy:

"One useful first-cut distinction is that between processing
gases/liquids and processing inhomogeneous solids, where mechanical
engineering issues loom much larger. I've been watching ideas for use
of _in situ_ resources since the 1970s space manufacturing
conferences, and have often run them by acquaintances who work in
mining, refining and process engineering.

Talk to them about Zubrin's scheme for automated processing of Martian
atmosphere, or even (farther along) "scooping" from gas giants, and
they say "maybe" (with caveats about how reliable valves, compressors,
etc, would have to be).
 Talk to them about handling kilotons of lunar regolith per month with
robots, telepresence, or a small handful of astronauts (who presumably
have a lot else to attend to)... using equipment that fits within
near-term mass constraints of lunar landing... and they laugh out
loud. With all due respect, Alex, have you ever actually witnessed the
start-up of a mine, cement plant, brick factory or the like?"

We're not just many steps, or even many major breakthroughs in AI, short of
that capability; we're many whole epochs of technology short of it.

I'm not trying to be a wet blanket - I think it can be done, given enough
time; I'm trying to figure out how to make the next major breakthrough. But
the difficulty of the task looks almost limitless, and the time available
looks anything but. That's why I think while studying existential questions
of the distant future is interesting and fun, the important problem facing
our generation is how to make progress _fast_ enough, while the window of
opportunity is still open.

This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:56 MDT