[ExI] matrioshka brains again, was: RE: Symbol Grounding

[Keith Henson]

I have been throwing rocks at this concept since Perry Metzger came up
with it.  (I found the original post and my response on a floppy
disk.)  There are two reasons it fails, heat sinking and distance
(speed of light) that slows computers to a crawl.  Despite my low
opinion of the idea, I got tagged as having originated it.

Keith

On Sun, Apr 23, 2023 at 10:14 PM spike jones via extropy-chat
<extropy-chat at lists.extropy.org> wrote:
>
>
>
>
>
> From: extropy-chat <extropy-chat-bounces at lists.extropy.org> On Behalf Of Jason Resch via extropy-chat
> …
>
>
>
> >…We could consider that it would have some self-healing mechanism, perhaps nanobots, or some time of liquid material that can be manipulated into a shape as micrometeorites ablate the material. Consider how trees can regrow leaves as they are lost…
>
>
>
> Ja, interesting ideas, but… at the time (2011) I set as a ground rule using existing technology only.  This was something we could design and make with 2011 technology entirely.  I wasn’t at the time suggesting anything we didn’t already have.
>
>
>
> In retrospect, I am not entirely certain I correctly accounted for coronal mass ejections.  Over the lifetime of our satellites, only a few years, it doesn’t matter, ionizing radiation doesn’t matter much, but over 1000 years… I need to ponder that some more.
>
>
>
> >…As long as each node on average can generate more energy than it takes to rebuild and replace that node before it's mean time to failure, then they can be self-replicating and self-sustaining…
>
>
>
> The 2011 version of the M-Brain had a separate orbiting node factory.  In that 2011 pitch, I didn’t offer any details on how that works (because I didn’t know any then (and still don’t.))
>
>
>
> …
>
>
>
> >…Electronics can be hardened to tolerate such things. There is ECC RAM, for example (using a few extra bits to provide error-correcting codes to detect and recover from random bit flips)…
>
>
>
> Ja we use that now, but what I don’t know is if there is long-term degradation that even now isn’t known (because it is irrelevant for satellites with a 10 yr lifespan.)
>
>
>
> …
>
> My strawman design had nodes spaced at 1 meter, so inherent latency would be 3 nanoseconds.  Thx for the sanity check Jason: no sanity was detected.
>
>
>
> >…I miscalculated as well, I should have said 0.3 ns based on my assumptions. ;-)
>
>
>
> Ja well we can cut each other some slack.  My goof was sillier than yours, because we yanks know that a light nanosecond is pretty close to a foot.
>
>
>
>
>
>  …
>
>
>
> >…I've heard this rule of thumb, and it seems to ring true: a computer will seem sluggish if it takes the CPU(s) more than 1 second of time to traverse all the memory in RAM…
>
>
>
> Before I can do much more with this idea, we need to firm up a mission.  ChatGPT introduces new possibilities I hadn’t considered, but I didn’t really expect anything that would interact with humans in realtime (still don’t.)  What I had in mind is a mostly independent ring of a trillion processors spaced at about a meter at 1 AU.  (Oh right, the bad news is we must assume away the earth and moon in this scenario.)
>
>
>
> >… Although I should clarify: I think when you said 256 GB, you are referring to non-volatile memory, which is more for long-term storage, rather than RAM which holds running programs…
>
>
>
> Jason in retrospect, I need to rethink everything in terms of long-lived processors.  We currently have no expectation of processors living for more than about 10 years, and seldom get that much out of a processor.  But an M-Brain needs to have a design which is adequate for the long haul, even if not optimal.
>
>
>
> When you think about it, we couldn’t make an M-Brain before now because processors change too quickly (and still are changing too quickly.)
>
>
>
>
>
> …
>
>
>
> >…Neural networks are implemented today as multiplications of huge 2D matrices holding floating point numbers. Graphics cards are well suited to this operation, far more so than CPUs are, which is why running deep neural networks benefits greatly from having a fast graphics card…
>
>
>
> OK this goes quickly outside my areas of expertise.  I can do the orbit mechanics stuff and the basic design is my own (Robert and I never did agree on it (mine was orders of magnitude simpler than what he was proposing 20 years ago (because it has no active cooling system.))
>
>
>
>
>
> …
>
>
>
> …
>
>
>
> >…Not to derail this project, but have you looked into the potential of using small blackholes as power plants? ( https://www.livescience.com/53627-hawking-proposes-mini-black-hole-power-source.html )
>
>
>
> >…I think it is promising for a number of reasons…
>
>
>
> OK cool, but as a parallel project, start thinking about a type of M-Brain subset, nodes which are manufactured on earth, carried into orbit using all current technology, let the fly around and operate as a proof-of-concept.  We don’t even need all that many nodes for that.  A few hundred would be plenty.  We want to let them adjust orbits, communicate with each other, verify solar radiation pressure and other really small forces can be controlled.
>
>
>
> For instance, all the orbit maneuver stuff is done entirely with solar radiation pressure, so we need to demonstrate that the whole idea isn’t wrecked by solar electrons causing these guys to take on a negative charge from having protons pass thru while electrons are captured.  This might be in my notebooks already: at what point does that net negative charge start to push the nodes apart with a force comparable to the radiation pressure?  I vaguely remember calculating that, but I don’t know how (or if) I ever convinced myself that these nodes wouldn’t gradually collect that much charge.
>
>
>
>
>
> …Robert thought so, but I fear that he persistently failed (or rather he flatly refused) to take into account something important: the thermal gradient.  I worked for a while on estimating that using Bessel functions, but eventually gave up on that approach because it was too easy for me to punch holes in my own reasoning.
>
>
>
> >…Can this be resolved by just making the layer very thin?
>
>
>
> Imagine a star with a diffuse foggy… well ok… fog, surrounding it, that is translucent: about half the energy gets out.  Make this fog envelope go from 2AU to 5AU (oh, right, bad new for Mars, it hasta go too.)
>
>
>
> Now we can estimate the temperature of the fog at 5 AU and it is very acceptably cool: just take the energy output of the sun and assume it radiating energy from a sphere of 5 AU with the old black-body radiation using the Stefan-Boltzmann equation.
>
>
>
> With that model, we get the 5 AU “surface” temperature, but what about the temperature at the 2 AU inner “surface” of the fog?  How do you calculate it?  My Bessel function approach was suggesting it would be hot in there.
>
>
>
>
>
>  …
>
>
>
> >…I think we can calculate what the temperature of the ring would be at 1 AU using the Stefan-Boltzmann Law. Incident solar radiation at 1 AU is 1360 W/m^2. In order to reach an equilibrium temperature, the ring needs to radiate 1360 W to stop increasing in temperature. According to the Stefan-Boltzmann Law, we need to solve for T in this equation: 2*(5.7603*10^-8)*(T^4) = 1360, note I use 2 here because the ring has two sides to radiate from, one facing the sun and one facing away from the sun... Jason
>
>
>
> Sheesh, if I had read the rest of your post instead of writing my way down to here, I woulda seen you already did what I was fixin to do.  How embarraskin, oy vey.
>
>
>
> Jason I am tired, let me rest and tomorrow I will finish with my thoughts on this.  Cool I am glad someone is taking a new interest in M-Brains, as I am because of the ChatGPT application.
>
>
>
> Tomorrow, we calculate!
>
>
>
> spike
>
>
>
> _______________________________________________
> extropy-chat mailing list
> extropy-chat at lists.extropy.org
> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat