[ExI] mbrains and latency

[spike at rainier66.com]

-----Original Message-----
From: Keith Henson <hkeithhenson at gmail.com> 
...
>
>> Dust-sized nodes can do station keeping, if they orient themselves with differential reflection.

>...I am not sure they can do anything useful.  An atom is around a nanometer, there are 1000 nm in a micron, so there could be a billion atoms in the cube.  I don't have Drexler's work in front of me, but that does not seem like enough to build a rod logic microprocessor much less a PV power system, transmitter/receiver, sail reflectors, etc.  A gigabyte of memory would burn the whole atom count...

Ja, but we are not locked into the micron scale nodes, even if we want to use that to explain Abby's star light dipping.  Even going to just ten microns, still dust, and that gives us a trillion atoms.  But we are not locked into a gigabyte of memory either.  Keith you and I remember the days when our HP programmable calculators gave us 969 programming steps with some of that having to be partitioned into variable storage.  I did a lot of cool programming with that little thing.

>...I don't know about the thermal problems either.  Might be ok, but I don't know how tiny objects radiate...

In very general, smaller is better.  As an object scales down, the surface area to volume ratio increases.

I wrote earlier about a trillion nodes communicating with others within a millisecond but I meant microsecond, with the entire teragroup communicating higher-level signals with another teragroup a millisecond away.  So a typical spacing between nodes would be a millimeter or so.  But they don't just send single bits to each other, they send small-ish data packets.

Another way to look at it: the human brain consists of a few billion neurons which have a number of inputs but don't individually calculate anything.  They just fire or not fire, depending on the input and the chemical environment.  A current computer network consists of a few million nodes, with each node performing a large number of very sophisticated calculations and sending each other very complex data packages.

The MSocieties then are somewhere between those two in a sense: the nodes are not nearly as sophisticated as a cell phone, but more sophisticated than a neuron.  A member of an MSociety, call it an JBrain (reaching back to terminology from 20 years ago coined by I think Anders) consists of a trillion nodes within a microsecond of each other doing a small amount of calculation and sending other nodes in that JBrain data packets of a few hundred or a few thousand bits.


>...Other than it being an idea from long ago, what advantage does this have over a solid structure?

Keith

It doesn't need structural strength, it isn't damaged by orbiting debris and meteoroids.  It can even be constructed from orbiting debris left over from an earlier pre-nanotech society suffering a Kessler event.  If so, it would be the only example I know of a way to recover space access after a Kessler catastrophe: nanoparticles disassemble the debris and construct dusi-sized nodes which go on about their business anywhere they want in orbit.

If a meteor from outside comes tearing inward, it would punch right thru a cloud of nanodes, and take out those of course, but the others which escape the collision would go about their business as usual with nothing analogous to a memorial service for their suddenly missing teammates.  Nanodes don't get emotional or have friends.

spike







> >...So a medium sized asteroid would provide enough material.
>
> >...Please check math.
>
> Keith
> >
> >
> > spike
>
> Keith your numbers look like they are in agreement with calcs I have done in the past.  I might dig out the old notebooks and look at my previous work once again.
>
> spike
>