COBE Data, MBrains & Ander's Razor [was Re: Metastable universe]

Robert J. Bradbury (
Wed, 27 Oct 1999 11:00:55 -0700 (PDT)

On 27 Oct 1999, Anders Sandberg wrote:

> I wrote:
> > Our "scale" for the variation in the CMBR is so "gross" at this point
> > it wouldn't take very big SI congregations to make things appear slightly
> > warmer in one direction than another direction.
> Hmm... If I remember the Cobe data right, the variations are rather
> large, several degrees across in the sky.

That is my current understanding as well. However think about how they collect the data -- I point a dish at a region of the sky all of the radiation coming from that region of the sky gets measured by the dish. It doesn't matter whether the radiation is coming from a point source within that region or is a proper average for the entire region, I still measure the same amount of radiation. The only way I can determine if its a point source or a uniform background is to have a very narrow focus and sweep it across the region.

It is similar to measuring binary stars. You normally measure the combined radiation of both stars. Only if you up the resolving power or observe for a long time looking for oscilations do you detect brightness variations that show what you are seeing is not a single object.

I doubt very much COBE would have "scanned" the sky since this requires much greater pointing accuracy and a much longer mission. It makes more sense that it would have done "point & measure". I think all we can say is that the measurements seem to indicate regions of space that have similar temperatures.

> This means that the
> congregations have to be very large (otherwise, they have to be close
> to us, and we should both see effects of them being inside our galaxy
> or local group (= anisotropies in the distribution), and a certain
> level of granularity would be expected).

I think about this and the problem of galaxy "walls". What if in some locations SIs get "lucky" and evolve early. They are going to use up all the resources in their galaxy and then spread out from there. [Robert F. has done some estimates that show that SIs can consume the resources of a galaxy *very* quickly.] Now interstellar distances aren't very limiting but intergalactic distances *are*. What you will see is regions of space where the SI "locusts" have consumed everything surrounded by the regions that they are slowly expanding into.

Eric D. once told me he had spent some time looking at images of galaxies for signs of expanding civilizations (dark regions). The problem is that you can "eat" galaxies relatively quickly. So galaxies exist primarily in a "natural" state or a "consumed" state. Eric may have missed this because he was thinking too "small and slow" (a rarity for Eric I'm sure :-)).

If there is a correlation between the "warm COBE regions" and the dark regions (without galaxies) then you have a very *interesting* situation. The nice thing is that this is a testable hypothesis once we have big databases with the locations of all of the galaxies and better resolution regarding the microwave background.

> If these concentrations are
> very large, then we have the same problem of causal connection again -
> how could the SIs have become similar over vast distances?

Intergalactic expansion. The reason we still see galaxies is that probability of the development of early SIs (i.e. rapid evolution of intelligence) is low. While I can't accept that it is difficult to evolve intelligence in 12 billion years (given our history as the only example), I can accept that it is difficult to evolve intelligence within say 3-5 billion years after the big bang. It probably isn't easy since the early universe is probably a more dangerous place and it takes time to accumulate the elements necessary for life. As the universe ages the probabilities for the development of intelligence increase. But once in a while in a rare galaxy intelligence could develop early. We would now be looking at space after those civilizations had evolved for 5-8 billion years. Given the scales of the galactic walls (100's of millions of light years), I think I can make an interesting case for "locusts".

> So the SI theory for the background radiation has the problem that
> either the SIs are close to us, but then they have to be so many that
> they look like a dense mist in the infrared, or more remote and then
> they run into the same causality problem of how they could have
> coordinated their temperatures given the finite speed of light and the
> short age of the universe.

Well the gravitational microlensing results estimate 200 billion objects orbiting our galaxy. If they "congregate" over time (to reduce communications delays) then that could explain some of the COBE observations. The temperatures (either around our galaxy or in the remote dark regions) are coordinated by thermodynamics. The background temperature might be entirely uniform. As you evolve to the limits and accumulate matter to fill the outer shells of the Matrioshka Brains, your temperature will drop lower and lower until it is just slightly above the background. You never operate your computers below 2.7 deg K since it is very inefficient from a thermodynamics standpoint. (Actually there might be some cases when some computational architectures could accept this "hit" but not many). I think Ralph M. and/or Eric had a paper where it said that to cool things to cryogenic temperatures (below your radiation temperature) you took a power hit of something like 30%. You radiate at slightly above the background temperature because thats all you can do and long term evolution drives you to that point.

> Another, perhaps bigger problem, is that if the anisotropies are due
> to SIs, the natural background radiation is going to be even smoother


> - and astrophysicists already have trouble
> with reconciling this level of smoothness with pattern formation.

But the assumption is that the pattern must be "natural". Turn it on its head and assume the pattern is "artificial". Assume the universe started out as *very* uniform. Assume in some "rare" places intelligence got an early start, then work your way forward from there.

> Actually, I would say in this case Occam is on my side ("Ho ho ho, now
> I got a razor..." :-), because the theory that the background
> radiation is mostly natural is much simpler than assuming SIs with
> some form of long-range coordination - that model introduces a
> tremendous number of new unknowns (like why some do and others don't
> radiate at a low temperature).

The *fundamental* question is whether or not you assume that the evolution of intelligence and the development of civilizations that evolve to the limits of physics is the *simplest* explanation. We (generally) *assume* that it is not. Our "simplest" explanation is that the universe is cold & dark & patterned & "dead". But the consequence of that assumption is that you have to assume as well that the evolution of intelligence is very very very difficult or the Fatality Probability for advanced ETC = 1.0!

Can you give me "simple" explanations for "very very very" and "1.0"?

Be careful how you hold that razor, I don't think most people have observed it cuts both ways!