Re: Humanoid Robots on the Mass Market

From: Franklin Wayne Poley (culturex@vcn.bc.ca)
Date: Tue May 23 2000 - 12:06:55 MDT


On 23 May 2000, Anders Sandberg wrote:

> Franklin Wayne Poley <culturex@vcn.bc.ca> writes:
>
> > I'll forward a posting on a proposal for a maximally automated
> > village. The question is, How far could we go with automating all standard
> > maintenance and chores like vacuuming, scrubbing, lawn mowing etc? I am
> > sure we can go very far if we are flexible with the initial design. Then
> > let's see how well the "genetic agorithms" do with improving the design
> > over time.
>
> Of course, by designing everything to fit the machines it can be made
> trivial (actually not even then, but anyway) to run an automated
> village. However, it might not be entirely pleasant to live in such a
> place since it will not be designed for human chaos.
>
> The problems one should think of carefully before setting up something
> like this is: what are you trying to achieve? Is this solution brittle
> and overplanned? Here in Sweden we have many buildings from the 30's
> created by progressive functionalists that had a lot of apparently
> great ideas about the house as a machine for living in. Unfortunately
> they ignored social realities, and all the helpful automation and
> systems they had added went unused.

Speaking of Sweden...have you been to Vasteras? I have read that they have
a heating system under the streets to melt the snow. Does it work
well? That is an example of an automated system to substitute for human.
Not unpleasant or chaotic (if it works). And how about "moving sidewalks",
another example of automation. In fact the whole internal transportation
system could be automated ... safe, pollution-free. Villages designed this
way could be linked by inter-village transportation methods, also safe and
pollution-free. People can leave their cars on the city perimeter.
 
> Genetic algorithms would likely turn the village into a nicely
> dangerous and chaotic place. After all, you can never be entirely sure
> the robots will be doing the right thing, and occasionally (likely
> quite often) unanticipated quirks will appear. When other robots
> encounter these quirks their quirks will also be activated, and the
> resulting co-ecolution might lead to something very much different
> from the efficient all-automated village originally enticipated:
>
> "Why is that half of the garden so unkempt? Well, our first lawnmower
> apparently had a mutation that made it avoid it. Then our neighbour's
> mower instead began to mowe it, and everything was fine for a
> while. Unfortunately it was rather "aggressive", running into any
> mower that came close, but it did a great job. After a few months our
> mowers learned never to go there. But then the neighbour's mower had
> an upgrade and only tended its part of the garden, and our mowers have
> long since bred out any genes that make them mowe that part. So we
> tried introducing some new ones, but they mowed my rosebushes, earning
> a lot of negative fitness, and the other mowers prevented them from
> going into the unmowed area since they had learned that overall
> fitness decreased when one of them got there and was ejected. So after
> a while we were left again with mowers that refused to mowe that
> half. Now we are planning to buy some lawnmower herder robots - we
> hope they can fix it."

It could get pretty nighmarish if the humans who start the GA's off don't
put them on the right path.
FWP.



This archive was generated by hypermail 2b29 : Thu Jul 27 2000 - 14:11:34 MDT