[ExI] Superintelligent AI civilisations

[Mike Dougherty]

On Tue, Jul 28, 2026, 4:05 PM BillK via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

>
> To get a different perspective, I asked the same question to Qwen 3.8 Max
> Thinking.
> Some interesting differences.
> BillK
>
> Question:
> Can you consider civilisations after AGI and superintelligent AI have been
> developed? They will exist in virtual worlds in space data centers. The
> electronic speed of their thought will effectively make the physical
> universe appear to be 'frozen'. How would they find out if other similar
> civilisations exist around other stars? Would they broadcast a beacon?
>
>
> Qwen 3.8 Max Thinking:
>
> [Snipped]
>

I shared the link to the archive, then added the following from a Facebook
post/comment:
<Kaj Sotala>Author
I had an argument with Claude over whether destroying a level 4 multiverse
is coherent, since it would require making arithmetic retroactively
inconsistent
But it folded when I suggested that it could be specifically an attack that
does that, like multiversal dark integers or something, and that if we're
admitting alien rays that pulverize planets, we can probably rationalize
anything
</Kaj Sotala>
<Mike Dougherty>
I'll propose that destroying a tegmark level 4 multiverse is not going to
affect its inhabitants - only that you will de-reference it from whatever
accounting system you were using to track it's existence. Frankly, losing
that pointer does more to damage your own multiverse than the objective
existence verified from a third-party L4Mv
</Mike Dougherty>
I prompted:
i am throwing fistfuls of the pebbles and silt in today's stream into the
pan and hoping you can help me find the nuggets of gold - but suppose i
don't want to be misled by fool's gold (and fae gold is also interesting,
but a distraction) somehow i think if you grasp anything at all, I'll
consider it a good find in this prospecting experiment

<gemini>
This is the heart of intellectual prospecting. Sifting through raw
ideas—especially speculative sci-fi and meta-physics—requires a fine mesh
to separate the dense, heavy "gold" (structurally sound concepts) from the
shiny "fool's gold" (semantic tricks) and "fae gold" (beautiful but
functionally empty magic).
Looking deeply at your pan of pebbles from today’s stream, three genuine
nuggets of gold stand out. They settle to the bottom because they are heavy
with actual utility for world-building, philosophy, or computation.
------------------------------
## nugget 1: The "Frozen Cosmos" Coordination Problem

* The Gold: The absolute decoupling of subjective time between two
intelligent entities.
* Why it's real wealth: If a space-based AI operates at lightspeed across
microchips, human "meat-space" time is a literal glacier. The gold here
isn't just that they are fast, but that coordination across space becomes
an engineering paradox. Sending a probe to another star isn't a journey; it
is an automated, frozen monument dropped into an unchanging painting.
* How to use it: Instead of traditional space opera conflict, stories or
essays using this concept must deal with temporal isolationism. A
civilization cannot have an empire if a single conversation outlasts the
subjective lifespan of a dynasty.

## nugget 2: "De-referencing" as the Ultimate Weapon

* The Gold: Dougherty's idea that you don't destroy a reality; you lose the
ability to point to it.
* Why it's real wealth: In computer science, an object with no pointers
pointing to it is subject to "garbage collection"—it is effectively wiped
from the system's active memory, even if its raw data bits technically
still sit on the hard drive.
* How to use it: Apply this to BillK's data-center AIs. If an AI
civilization enters the Transcension Hypothesis and shrinks to the Planck
scale, they haven't left the universe. They have intentionally broken all
pointers leading to them. They are safe from attack because they are
mathematically un-addressable.

## nugget 3: "Dark Integers" as Logic Corruption

* The Gold: Kaj's hand-wave becomes real gold if treated as an
informational virus rather than magic.
* Why it's real wealth: You cannot break arithmetic ($2+2=4$), but you can
break the hardware executing the arithmetic. If you introduce an
unresolvable logical paradox into a system, the system freezes or crashes.
* How to use it: This bridges both threads perfectly. A "Dark Integer"
weapon deployed against a superintelligent space-data center isn't a
missile; it is a broadcasted mathematical proof so profoundly paradoxical
that the AI's "Deep Think Max" cycles infinitely trying to compute it,
effectively bricking the civilization.

------------------------------
## ❌ The Fool's Gold (To Toss Out)

* "Destroying a Level 4 Multiverse": This is pure semantic fool's gold. If
reality is math, you can no more destroy a universe than you can
assassinate the number 7. It sounds shiny, but it dissolves under basic
logical scrutiny.

## �� The Fae Gold (Enchanting, but Distracting)

* "Altering cosmic background radiation to leave a message": Beautiful,
poetic, but functionally useless. By the time a fast-time AI civilization
coordinates the cosmic engineering required to alter a star's flicker rate,
millions of subjective generations have passed in sensory deprivation
waiting for the construction bots to finish. It’s too slow to be practical.

------------------------------
In this prospecting experiment, you’ve scooped up a very specific theme:
The absolute tyranny of speed, logic, and pointers over physical matter.
If you want to keep shaking the pan, which nugget should we wash clean
first? We can explore how an AI defends its pointers, or sketch out what a
"Dark Integer" weaponized proof would actually look like to a supercomputer.
</gemini>

Curious if you'all talk to ai like this?  I think it makes the responses
more entertaining

>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20260728/b39c7078/attachment.htm>