[ExI] AI computers and power in space

[Keith Henson]

BTW, Keith Lofstrom talked about a variation on this idea some years
ago.  Should have mentioned that.

Keith

On Mon, Apr 29, 2024 at 9:55 AM Keith Henson <hkeithhenson at gmail.com> wrote:
>
> John Clark posted:
>
> Apr 28, 2024, 4:54 AM (1 day ago)
>
> to extropolis, 'Brent Explore this gift article from The New York
> Times. You can read it for free without a subscription.
>
> In Race to Build A.I., Tech Plans a Big Plumbing Upgrade
>
> The spending that the industry’s giants expect artificial intelligence
> to require is starting to come into focus — and it is jarringly large.
>
> ^^^^^^
>
> My work on thermal power satellites got the estimated cost down to 3
> cents per kWh.  Without the 50% transmission loss, 1.5 cents per kWh
> is possible.  In addition to the huge radiator area, heeded for the
> power generation, it would require around half again as much area to
> get rid of the computation heat.  On the other hand, the design does
> not need a power transmission antenna.  Further, it could be placed in
> an orbit lower than GEO inclined more than 23 deg to avoid the shadow
> of the Earth and the thermal shock from the shadow.
>
> Thermal radiators are arranged as pairs of tapered plastic tubes
> filled with 20 deg C condensing steam.  The tubes make a planer
> surface facing solar north/south and shaded from the sun.
> Concentrator mirrors focus sunlight into boilers or PV.  Power
> generation and computation would happen at the ends of the tube pairs.
> There is an animation of a 1.5 GW radiator about 4:30 into
> https://www.youtube.com/watch?v=VEkZkINrJaA  That may be larger than
> any data center on earth.
>
> Data centers are already generating a fair amount of NIMBY pushback.
> This is a way to push computation off Earth.
>
> There is an extremely speculative safety reason to move AI off Earth.
> If the AIs ever went rogue and were already in space, they might leave
> planetary surfaces deep in gravity wells alone.
>
> Keith