[ExI] simulation as an improvement over reality

Eugen Leitl eugen at leitl.org
Thu Dec 30 13:41:36 UTC 2010

On Thu, Dec 30, 2010 at 03:46:41AM -0800, Samantha Atkins wrote:

> Sure. Someday.  But not first to attempt to gain cheap 
> and plentiful enough energy to keep our technological 

I mentioned the threshold to SPS to be PV contribution to total
at high multiple 10% across industrialized nations. Right now 
Germany barely managed 2%, peaking at 20% in rare days.

I doubt the exponential trend can keep up, since there are
no scaling laws to help you. Even if you can ink up km^2,
you still have to put them up, wire them up to the grid
and/or buffer the power.

> momentum up.  We need that long before we are capable 
> of deploying SPS systems in GEO and/or on the moon.

I agree. How quickly the need arises depends on ability
to transport energy over long distances (high-voltage
DC, hydrogen pipeline) or buffer them. Supercaps might
make a difference, but you'll probably need large-scale
cheap electrochemical energy sources. Availability of
either in time frame required is not guaranteed. 
> > 
> >> The reason is that we don't have the robotics 
> > 
> > I don't think it's a hard problem these days.
> > There have been a number of past bootstrap
> > plans, see http://www.molecularassembler.com/KSRM.htm
> > 
> Lets see.  4000 - 20000 metric tons of panels, 
> concentrators, antennae and so on for an SPS.  

In general you would want to use um-thin layers of
photosynthetic material (probably also doubling
as solar sais), and with phased-array microwave
radiators (capable of realtime beamforming to track
groundside rectenna arrays with smooth handover)
integrated into the bottom of same array.

The difficulty would be in fabricating these and
assembling them into packages which can be launched
via linear motor (with some rocket assist), and
can be aerobraked.

You would be probably easier to fabricate metal
and semiconductor-grade silicon ingots and insert
them into suitable Earth orbits, to be processed

> Assuming you can get it out of the gravity well 

Out of the lunar gravity well. Dropping them
into the Earth well is much easier, with

> (or mine and manufacture in near earth space) 

That might make sense, since making very large
extremely thin panels might require microgravity.

> then there is a huge need for workers / devices

Teleoperation. Even easier in LEO.
> of some kind to assemble and maintain all that 
> lot in GEO or on the moon.  We don't have 

GEO is too far.

> anywhere near that many astronauts and do not field 

Monkeys will stay on the ground. Teleoperation
platform can be handed over 24/7/365 for
groundside crews.

> them for space walks in GEO on an extended basis 
> in any case (much harsher radiation danger than LEO).   
> So having humans do it is not going to cut it, not 
> in the next decade or two at least. 

I think space monkeys is going to become an even
rarer species with each coming decade.
> >> or huge (and expendable) human astronaut population 
> > 
> > Space suits don't really work, so why not use
> > teleoperation? Even NASA has wisened up meanwhile
> > http://robonaut.jsc.nasa.gov/default.asp
> You can't beyond a certain range for anything requiring 
> relatively continuous adjustments.  But you can run more 

Yes, which is why I keep harping about the next rock
from here. 2.5 pingpong times are still doable with
trained operators, but they also give you incentives
to install local reflexes into the teleoperated system,
turning the human operator increasingly into a planner,
and the remote system as a (pretty dumb) executor of
a plan.


> nearly autonomous bots at greater distance.  But we don't 
> have hardly any construction bots fully proven out in space.  
> Some of the attempts have had serious problems such as 
> joints that freeze up if the power ever drops.  

The problems with UHV are pretty much fixed (though of
course you have to design for that). Working in abrasive
dust environments, with electrostatic charging, and also
in cryogenic, very cold environments (<90 K in polar
lunar craters) will pose some challenges.
> > 
> > And of course there's not much point in teleoperating
> > the robot from nearby pressurized vessel, if you could
> > leave the human at home. The Moon is close enough
> > for teleoperation,
> Not for a lot of teleoperated processes.  

There are not a lot of teleoperated processes
on this planet. There are very few humans hurtling
through the thermosphere, with very little time
spent outside. They're still using robotics arms
and you'll notice that the R2 will become a permanent
resident on ISS.

> If it is then why don't we have tele-operated rigs 
> running all over it now?  

Because we no longer can go to the Moon. You would first
need to expand DTN-capable Internet networks to lunar 
orbit (there are experimental Cisco boxes in LEO already), 
and be able to soft-land instruments and robotics platforms.
> For an SPS in GEO 

GEO is too far. And already too crowded.

> the break even size is around 2 - 4 GWs.  

I have not seen an EROEI calculation for a
lunar launch, but any self-rep system with a 
closure over unity is the ultimate free lunch.

> That is a very large space structure to 
> assemble and maintain. 

You would want as many small (>100 m) platforms
as possible in as low an orbit as possible.
You need 2-3 of them permanently in line of sight
for a smooth handover. You'll need hundreds to
thousands of these.

You would also use them to kill GPS, Iridium,
and satellite Internet providers. Later, cloud
services. And so on.
> Sure, although space robots have additional challenges.

> Molecular assembler?  Fine.  But Drexler, Merkle, 
> and Frietas seem to be predicting no less than 
> three decades to get to machine phase.  I don't 
> think we can wait that long to exploit near-earth space.

I see you haven't read the link. The full-text online
book is a literature review of self-replicators in general,
including space-based manufacturing.
> > At this point, it's largely a question of budget.
> > There's a second Moon race of sorts on, Japan,
> > India and China (and Europe) being the main participants.
> > 
> Personally I think the moon is the wrong target for 
> doing near earth space infrastructure.  There is more 
> variety of useful and lucrative material in near earth 

Material transport is to be avoided, initially. 
Material trasport to Earth will never amount to much,
in my opinion. There is an incentive to move heavy
industry and associated contamination outside of
Earth's ecosystem, and only deploy prefabricated
and fabricated components. Also, if we are at that
point, the human era is shortly over, anyway.
You'll notice that the ecosystem itself does little
matter transport, prefering to utilize materials
in situ.

The nearest (2.5 s relativistic pingpoing) rock is
the Moon. It doesn't have microgravity, which is an
advantage if you're learning the ropes.

After you've got nearly autonomous system and experience
in microgravity manufacturing, the whole solar system
is yours.

> asteroids more easy to get to.    No problem with doing 
> both but I think the bounty of near earth asteroids 
> is very seriously overlooked. 

I used to think that the asteroids were the way to go.
However, the Moon is the gateway.
> >> how you are going to create the ability to mine, 
> >> process, assemble and maintain all this infrastructure 
> >> and in what steps.
> > 
> > The basic idea of ISRU is to reduce mass transfer.
> > It's pretty obvious that you don't need to fabricate
> > control logic locally, and for e.g. PV panel production
> > the silicon (or other materials) can initially be
> > transported from Earth.
> At $2000 - $10000 / kg?  

Thin-film CdTe is some 10 g/m^2. So a ~kg buys
you around 100 m^2 of solar panels, if sputtered
on regolith glass. If you're bootstrapping
power plant capacity to power your manufacturing
capacity, any kg less reduces your bootstrap costs.

I'm assuming >100 kg packages semi-soft-landed
(300 g would do) on lunar surface.
> > We know there's a lot of
> > volatiles at the pole(s), which is incidentally 
> > one of the few places in the Solar system where 
> > ancient crater cryotraps sit right next to the 
> > peak of eternal sunlight.
> There are a lot of volatiles in your average near 

Well, there's a lot of hydrocarbons on Titan.

> earth asteroid and a lot cheaper to extract. 

We know it will be very cheap to extract, microwave
heating and condensing in adjacent polar cryotrap 
will do. 
> Volatiles for delta V to move the lot (or the 
> extracted materials you wish) to where it is needed.  

We need volates for wet industrial processes (e.g.
electrolysis and hydrogen for oxide reduction, 
and such) and maybe some oxygen/hydrogen to
make it to Earth's atmosphere.
> > 
> > So that looks like the best place to start.
> > 
> > The bootstrap won't be completely scripted,
> > since there will be always surprises, improvisations
> > and optimisations.
> > 
> We don't have the means to do this yet,

If we wait much longer, we'll never will.
> principally we don't have cheap enough 
> launch or the assembler/maintainers in 

Chemical rockets to make it to LEO and
plasma thrusters to make it to low lunar
orbit are all we're going to get.

So your only other variable to minimize
the total number of tranfers, and mass
trasfered. That is a function of 
telecommunication, computation, and

All of this can be done down here.
But for 1/6th g you can simulate the Moon
environment in lunar simulators, at a tiny
fraction of costs it would have cost you to
test-drive your gear under real conditions.

> space or on the moon.   They can't be 
> human in a timely manner in the quantity 

There's no need for humans on the Moon for
the foreseeable time, if ever. I do not
think there will be permanently manned
presence outside of the inner solar system
(Moon and Mars).

You want to go to the stars, you have
to lose this mansuit.

> needed.  Thus, robots - much better robots.   
> More than a few humans are too expensive 
> and difficult to maintain in space or 
> on the moon until substantial infrastructure 
> is already in place. 

Yes, robots. And then, robots which are

Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org
ICBM: 48.07100, 11.36820 http://www.ativel.com http://postbiota.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE

More information about the extropy-chat mailing list