From: Emlyn (emlyn@one.net.au)
Date: Sat Mar 17 2001 - 18:38:57 MST
I have (hopefully) clarified some of my points below; it appears my meaning
was obscured.
Robert wrote:
> You can perhaps "duplicate" or "grow" intelligence remotely but the
> return on investment seems pretty low because they can only communicate
> a small fraction of what they "know" with each other. Now, say you
> make the transition from we are now (where interhuman bandwidth is
> very low relative to internal bandwidth) to the posthuman state
> where inter-posthuman bandwidth is very high (I can "know" everything
> that you "know") -- are you going to want to return to the "human"
> state where you can't communicate much of what you know at all?
Maybe an MBrain SI would be curious to find out what another of similar
abilities would think of, after a given period. Assuming that one MBrain
cannot entirely simulate another (can we assume that?), then a seperate
MBrain, existing at a distance, potentially knows some incredibly useful
things. The case for forking new brains, which build from scratch using
nanospores and information parcels of whatever kind, seems pretty good to
me. Can two MBrains do considerably more than one? What is the cost of
sparking off a new MBrain on a distant star? Just about negligable, I'd
think, meaning that the perceived benefit doesn't need to be too high; this
is assuming that there is no perceived risk, which I take it that Robert
postulates there is not (ie: No MBrain wars).
>
> The information cost of leaving the "hive"-mind is very high.
> I suspect most would choose never to do it.
But you might send out some matter to create a new mind; no one is leaving.
>
> > There is a discussion of EM vs matter. How is it possible to expand as
EM?
> > There's no receiver, no way to kickstart a replicator at the other end.
>
> Sure there is, you construct a probe with enough matter to hold the
> MBrain designs, and enough nanoassemblers that some of them are still
> functional after traveling through interstellar space. You then launch
[snip]
I should probably have said EMR here. I am specifically not talking about
transmission by mass; I'm talking about some frequency of electromagnetic
radiation.
> > So are we (I'll be there) not entirely limited to *initial* transmission
> > using matter? Damien asked the same question on list a while back,
actually,
> > and I don't remember anyone answering "you can use em as so ...."
>
> The BIG problem that I wasn't clear about earlier is that given the amount
> of "information" that you rapidly get at each end (> 10^40 bits
potentially)
> one *cannot* communicate any significant fraction of that across
interstellar
> space using electromagnetic radiation (using any of the "popular" ideas
> in SETI, OSETI, etc.).
I guess I was saying that, even if you could transmit a decent amount of
information over a long distance using EMR, who/what will receive it at the
other end, if this is the initial expansion (ie: you've never been able to
send matter or detect other beings at the destination point).
>
> Major Consequence:
> You can *Never* read Encyclopedia Galactica if it transmitted
> via electromagnetic radiation.
>
> Now, I've thought a bit more about this since I first let my fingers fly
> and you may be able to get a bit closer using wavelength division
> multiplexing at light and UV frequencies but I'm doubtful it will
> solve the problem entirely.
>
> The only way you transmit information around the galaxy is as high
> density "matter" packets.
>
> > (There is a possible answer to that, of course; you send out an SI seed
> > blueprint in a form that a low-tech civilisation could intercept and
use.
> > Very dodgy in many ways, though: Is there life out there to receive it?
> > Who's going to receive it; low tech friend or scary higher tech SI
gobbling
> > foe? etc etc)
>
> I've addressed this in previous discussions. You never send stuff out
> to regions beyond your "light horizon" because you can't guarantee
> what will be there when it gets there. The implication of this is
> that any colonization or forking off children is likely to be only
"local".
>
Could you talk more about this? I don't know what "light horizon" is
(apologies for ignorance).
> > So it seems that sending modulated EM is not useful, except as follow up
> > signals to an initial spread of spores. Except that Robert tells us it
is
> > too wasteful.
>
> I think you can transmit small stuff like a nanobot design (you do have
> GHz data rates) but you *do not* transmit significant fractions of your
> knowledge base. If you want high-bandwidth communication you are going
> to be hurling asteroid sized "data packets" back and forth. They are
> probably quite "pretty" packets as Mike's recent picture showed.
>
Hmm... you might want a bit of security on those APIs. Capturing and reverse
engineering someone else's nanospores could be pretty useful, if you can't
be bothered sending out your own... capturing their data packets would be
pretty useful, too.
> > My second point is, given that you've already got matter out there in
some
> > form, you've had time to put a bunch of relay stations *out there*,
ready to
> > relay your signal. A big long (long!) line of them could take a signal
from
> > one star to the next, without anywhere near the losses or a super death
ray,
> > could it not?
>
> Yep, the interstellar pony express does seem feasible except when they
> have to ride through Amara's clouds of dust. The ablation in those
> probably makes them take the long way around.
>
What I have been unclear about here, is that I mean lots of physical
repeater stations, transmitting *EMR* from one to the next. The distance
between each is short, and they span the distance from one star to the next.
Do they need to move? Maybe they can orbit in a giant, wide orbit with the
source and destination stars at the focii of the elipse? Is the closest
point between star and elipse edge then close enough for initial & final
transmission?
To fix that problem, you could probably have secondary local orbiting rings
of relay stations, which orbit only one star, and which come close to the
larger double-star focussed orbit. Transmissions go to your local relay
stations, pass around the the closest point to a relay station in the long
double-star orbit, transmit to it, then the signal travels around the long
orbit to the second star, into it's local set of relay stations, then in to
it's local receiver.
> > What they are powered with is an interesting problem (the energy they
receive?),
>
> You want to recover as much as possible. Magnetic braking of the incoming
> data packets should get you back most of the launch energy and you simply
> recycle the matter for outgoing packets (or perhaps you have clever ways
to
> do the readout nondestructively, then simply do any necessary rewriting
> flipped bits and send the packet on to the next station).
With EMR transmission between stations, I am wondering if the collected
energy can to some extent be used as signal & as a power source.
>
>
> > My third point (I've a disordered mind) is that the cost of EM
> > communication might be high, but the speed may justify the cost.
>
> Yes, I agree of course. The question is what information is so valuable
> that the communication justifies the expense?
> Transmission: "Hey my sun just went nova!!!"
> Reply: "Really, we predicted that 3 million years ago..."
Probably it's information to do with the actions of peers. Stuff that you
cant predict.
>
> > Information is important to us; maybe it'll be important to giant
> > spaceborn brains? Timelines often makes the difference between
> > information and noise. Waiting for your sub-c message-in-a-bottle
> > may not be an option, if others are communicating at light speed.
>
> If you want to communicate large volumes of information you only do
> it with EMR if you are in *very* close proximity (sub-lt yr distances).
>
> You can use IR (or radio) to transmit through the dust clouds that
> your information transport packets can't go through but the data
> rate is going to be pretty darn low.
>
> I think this may resolve the Fermi Paradox -- SIs only fork off
> children when they get *very* close to something (solar system,
> brown dwarf, etc. that can be rapidly turned into something
> that can "hold" a download of >10^40 bits. You have to be
> close enough to it so that you can align billions of lasers and
> receiving antennas such that you don't waste the energy involved in
> the information transmission and can get the foundation data for
> the child across the void while you are still in close proximity.
>
> This doesn't happen very often because stars don't get that
> close that often and navigating an MBrain into close proximity
> requires the expenditure of a lot of energy and possibly mass
> and takes a very long time (millions of years minimally).
>
> Now, this would "predict" that the center regions of galaxies
> should grow dark more rapidly than the outer regions of galaxies
> (more close enounters providing more forking opportunities).
> I don't believe we are observing that. Perhaps SIs close to
> the centers of galaxies have more to worry about like how
> to minimize the cost of staying out of the black hole over
> the very long term. They can't afford to worry about offspring,
> in large part because they are more concerned about their long
> term survival. Now *this* gets very interesting because it implies
> one possible path civilizations might follow is "star hopping"
> *away* from the center of the galaxy. You don't "fork" at all --
> you transfer yourself en-mass to stars that are "outward" bound.
> I'm pretty sure that I wouldn't want to be on the MBrain Titanic
> as it gets sucked into oblivion. Now, this *would* explain
> why there seems to be more "dark matter" in the outer regions
> of the galaxies -- those are the neighborhoods where one
> can expect to survive the longest. Long term, you are looking
> for a travel path that minimizes getting sucked into a black
> hole or colliding with another star and allows you the greatest
> opportunities to pick up elements that are in short supply
> (and possibly allows forking opportunities) until you can make
> it to the "safe" region of intergalactic space.
>
> In case this isn't clear, the implication (if this is correct)
> is that advanced civilizations do *not* expand at a velocity
> remotely approaching 'c'. They expand at the rate at which they
> can arrange convenient (read inexpensive(?)) interstellar 'forking'
> opportunities. Those probably only occur once in a blue moon.
>
> Robert
>
>
Robert, there is something about your theory that comes to bother me, in the
above paragraphs. I'll try to explain it.
When you ascribe motivations to MBrains, they are always about survival and
energy minimisation. As a consequence, they don't communicate, they don't
move around too much, they spend energy on things like escaping a black hole
that might suck them up in a couple of million years.
This seems to be an entirely purposeless existence. These brains don't do
anything, they just try to survive. That's fine, there are lots of organisms
that do that. However, MBrains are I think the maximal resource users of any
organism that I can imagine. These things swallow whole stars!
If your goal is survival, there are lots of ways of doing that using a
damned lot less energy and matter. Firing off zillions of nano spores in all
directions, containing all your recorded information, might be a start.
Becoming a spaceborn insect plague that spreads across the universe.. that
might work.
What your MBrains actually do with themselves, would seem to be achievable
using an old Z-80 chip. So why do they need such gigantic brains?
Basically, I think that you would need a reason to build such enormous
brains. Pure survival is not that reason; it seems quite clear that the goal
of survival could be accomplished in a much superior fashion, using the
simplistic methods you've described (run away from dangerous things and
don't talk to anyone) by using far less (far less!) resources, and being
smaller.
So, an MBrain's primary motivation is either not survival, or else it is
survival, but in a way so extraordinarily non-obvious that you require a
whole solar-system worth of computronium to figure it out.
I think this point must flow through all reasoning about MBrain actions. If
we cannot know what MBrains are motivated by, we cannot know how they will
act. Surely, they will be prepared to consume resources to come closer to
their final goal. So this idea of floating quietly in space, desperately
conserving local matter, drifting aimlessly toward the heat death of the
universe... it just seems a bit silly.
Emlyn
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:06:28 MST