[extropy-chat] [Bulk] Re: Forbes Magazine on Robotics

Keith Henson hkhenson at rogers.com
Tue Aug 22 01:45:46 UTC 2006

At 03:40 PM 8/21/2006 -0700, you wrote:
>On 8/21/06, Keith Henson <hkhenson at rogers.com> wrote:


> > This has been discussed here and on the SL4 list without a definitive
> > resolution.  Smaller is better for communications, but engineers worry
> > about cooling.  Other places I have mentioned sinking uploaded societies in
> > the deep ocean for cooling.
>The deep ocean is extremely hot relative to the best engineered
>cooling solutions (liquid helium for example), and I don't see any
>reason why this sphere can't be suspended in the vacuum of outer

This is something Eric Drexler and I investigated at length back in 1979 
when we wrote a paper on gigantic radiators for use in space.

It's not easy to dump waste heat in space.  If the radiators are operating 
at room temperature, they can get rid of about 1/4 kw of waste heat per 
square meter.  And large radiators have diseconomy of scale.  If you make a 
radiator 100 times as large it, it weighs ten times as much per kw.  It is 
really hard to beat a planet covered with liquids and gases for a heat 
sink, a point Poul Anderson wrote a story around long ago.

>Also keep in mind reversible computation:
>And Robin Hanson's excellent paper on reversible agents:
>When a civilization has control of every atom, waste heat need not be

See Eliezer's comment.  He has it right.

>Waste heat is defined as atoms getting out of place and
>bumping into each other.  In a civilization with atomic control over
>its own matterspace, you can simply catch the atoms and reroute them
>before they bump into each other destructively.
> > The problem is *when.*  If it is clear that humankind will leave flesh
> > behind in ten years there is no need to embark on an energy project to feed
> > 6+ billion people.  If uploading takes 30-40 years then we need a massive
> > energy source coming on line to displace the current ones.
>It could happen very quickly with recursively self-improving

The operative word is "could."  It may go fast once it starts, but we have 
no idea of how long it is going to take to get to the point it takes off.

> From the perspective of the gaian biosphere's
>evolutionary processes, "this human thing" exploded everywhere in the
>blink of an eye.  Intelligence building intelligence could easily do
>the same.
>It is unknown how long this will take though.  So let's cross our
>fingers that ITER goes online smoothly.  And that we build cars and
>power plants that can take ethanol.

Ethanol is just barely a positive energy return.  If they didn't cut the 
cane by hand in Brazil with extremely low cost labor, it probably would not 
work at all.

> > Given that war memes do well in societies seeing a bleak future, it might
> > be worth starting even if you think uploading will come before a full scale
> > space elevator could be completed just to keep the war memes down.  An
> > unfriendly AI emerging as the result of desperate war research is not the
> > stuff of pleasant dreams.
>I'm not too excited about advanced weaponry and UAV swarms being mass
>produced using Phoenix nanofactories, neither.

If you don't want war, then you need to expand the economy fast enough to 
get ahead of the population.  Energy from space would let you do a whole 
lot of things.  For example, the cost of aluminum is mostly electricity.

Keith Henson

More information about the extropy-chat mailing list