[ExI] Computronium planet.

[Keith Henson]

On Wed, Oct 17, 2012 at 5:00 AM,  Anders Sandberg <anders at aleph.se> wrote:

> On 12/10/2012 08:35, Keith Henson wrote:
>> The radiators would depend on what they were doing. Used for fairly
>> slow access, a planetary scale device might not need to much cooling.
>
> Yes, but even a tiny energy usage per cubic meter would get hot fast. A
> sphere with radius R producing P Watts per cubic meter will need to get
> rid of 4 pi R^3 P/3 Watts. It can radiate from the surface, 4 pi R^2
> epsilon sigma T^4 Watts. That gives an equilibrium temperature of [R P/3
> epsilon sigma]^(1/4). So if R=6e6 meters,  P=1e-6 W/m^3 the temperature
> will be around 100 K. For a milliwatt it will be 570 K - good for
> cooking. A Jupiter-sized version would be just as hot with a microwatt
> performance.

Agreed.  Thinking of information storage without a lot of access.

>>   It would be an interesting place, dark, faint spill of light around
>> the sun blocker, land areas covered with rectennas, fast uploaded
>> civilization in the depths heating the the oceans to steaming,
>> freezing cold rain falling everywhere, vast rivers running off bare
>> rock continents. Inside the simulation conditions could be as nice as
>> anyone wanted, but the underlying reality would be stark, worse than
>> Mordor. Be hard to detect though, since the shading disk would
>> probably be circular.
>
> Sysop Sauron was watching the cooling indicators of Mount Doom. "We are
> seeing a spike in server latency, what the heck are those heavens doing?"
> "They are apparently running a cross-civilization marathon game. Should
> be over soon when the NPCs have evolved sentience."

Small and fast means a lot of waste heat in a small space.  Liquid
cooling seems like a good idea to take the waste heat from hot
computation nodes to where it can radiate to space.  I wonder what
weather patterns you would get with very little sunlight and most of
the heat coming from the oceans?  At what light level do the plants
die?  Are Robert Kennedy's "Dyson Dots" and my power sat proposal on
the evolutionary pathway to a dim earth?

But this brings up a point.  What is the optimal distance from a star
to do the most computation?  The further out you go, the more
collector area you need to power the civilization.  Closer than some
point gives you more power from the sunshade than you can radiate.

Is a planet useful?   I think it does for a communicating, fast
civilization.  But is it more useful than a fog of computronium?  At
only a modest million to one speedup, stuff on the far side of the
Earth's orbit will be over 2000 subjective years away just from speed
of light.

Keith