[ExI] Tabby's star

Keith Henson hkeithhenson at gmail.com
Sat Dec 10 00:36:21 UTC 2016

It's going to take more effort than I thought to make sense of the
weird data we have from Tabby's star.  But here is a start.

Shame this was not found back in the early 90s when the consequences
of nanotechnology and AI were hot topics on this list.  They are the
lynch pins of the singularity.  I might add that for engineering
reasons (thermal and speed of light) I disagreed with them on a number
of points, in particular on Jupiter brains, and Matrioshka brains.  I
seem to have been wrong, apologies to Perry Metzger and the shade of
Robert Bradbury.  If we are looking at a post (local) singularity at
Tabby's star, they built something 7 times as wide as Jupiter.  My
guess is that it's a self-replicating, Google type data center.  My
concerns about signal delay apparently were not an overriding

Apparently they traded off low temperature operation for latency.
Jason Wright thinks the blocking objects are below 65 K, Keith
Lofstrom makes a case that solid state computation optimizes below 60

Assuming the blinks at KIC 8462852 are from constructed objects
crossing the star, the first problem is to figure out a motivation
within known physics for building them.

I have been concerned with heat sinks in space for close to 40 years.
Waste heat concerns are ubiquitous in engineering.  Getting the waste
heat out is what capped rising clock speeds for microprocessors.   If
you have a star, especially a hot F type, getting rid of waste heat,
especially at low temperature, could induce some very big radiator
construction projects.

I wrote about heat sinks and clock rates here:


If the 22% event is a disk and went right over the solar disk, the
orbital velocity would be around 10.5 km/s (which is what Dr. Boyajian
estimated) and the orbital radius around 11.5 AU.   For that star the
velocity puts it at 11-12 AU.  Saturn in our solar system is about 9.5
AU. We might see it go by again in about 30 years.

The 15% event could be from a much larger object.  Assuming a disk,
the chord of the disk blocking the star took ten days to cross the
star.  If it is at roughly the same distance from the star, the chord
is about 9 million km long.  That would make it about 35 times larger
in diameter than the star.  Alternately, it could be much smaller and
much further out.  For half the velocity it would need to be about 45
AU out from the star.  It would still be enormous.  At that distance
we would not see it again for another 260 years.

If they are running the radiators at close to liquid He temperature, I
don't think we are likely to detect them.  I wonder how low a
temperature the radiators can go and still be detected?  (If they are
radiators, of course.).

The evidence from the past observations somewhat confirmed by the
Kepler data is that the light from this star has dimmed by 15% over
the past century.  That would mean that 1500 years ago they were still
in a dynamic growth phase.  To see the blinks from large light
blockers, we need to be close to the ecliptic of that system.  How far
north and south of the ecliptic the objects soaking up the light that
comes toward us is unknown, but even a degree or so would still be an
awful lot of energy.  If what's catching the light is going through a
thermal cycle, then they need to either be so far out that the
radiators are below 65 K, or they need to be of a kind that radiates
the waste heat local north and south away from our line of sight.

Conceptual examples that radiate directionally have been designed.


The missing IR flux (WTF) is not consistent with this distance for a
natural object in thermal equilibrium, not even a disk.

I suspect (if we really are looking at alien megastructures) that they
are shorter on materials than they are of energy.  The blinks we see
at Tabby's star seem to make sense in engineering terms.  If you want
to maximize the amount of computation, given limited resources (and
resources are always limited) then going out from the Sun (or star) is
a good idea to get the temperature down.

This leads to testable predictions.  I am kind of surprised at the
size, but it's possible the computational core(s) is/are small but
being cooled by a radiator on a size comparable to the star.

How small the cores are depends on how fast or slow you want to have
the universe run around you.  Edge to edge, the 22% dip is a million
km in diameter.  It would take light 3 second to cross that distance.
If the inhabitants are running at human speeds, that's too slow for

What on (or rather off) Earth could they be doing with computation on
that scale?  Well, you know about the simulation hypothesis?  That
would sop up as much computation as you want.  "The Clinic Seed" story
worked around the edges of uploading humans into a simulation in the
"boiling a frog" mode.  Accelerado toward the end has the Vile
Offspring generating simulations of everyone who left a written
record.  Consider that when posting.  :-)

More information about the extropy-chat mailing list