[ExI] ai comments on m-brains

Tue Oct 21 20:39:57 UTC 2025

Cool thx BillK.

What I found interesting is that this passage that BillK posted is the second round of inquiry.  ChatGPT came back with a set of arguments against an M-Brain, which I refuted.  It then came back with something humans seldom do: a modification in its “thinking.”

The Gemini 2.5Pro (my apologies to Gemini for calling it ChatGPT (but note that ChatGPT has first-to-market advantage of becoming the generic name for LLM-based AI (analogous to how all facial tissue is called Kleenex regardless of brand))) contributed something important, even if comically understated.

My contribution to the M-Brain discussion: take an engineer’s approach rather than a physicist or mathematician’s approach.  The mathematician, the physicist and the engineer just think differently from each other.  The mathematician is super-cautious, carefully identifying every assumption, disciplined, careful.  This approach fails for the M-Brain.  The physicist carefully derives equations to use the mathematician’s toolkit he so generously provided, making everything general.  But this approach also fails (which is why we have no general equations describing the thermal characteristics of an M-Brain that work right.) 

The engineer is kinda morally questionable, kinda loose and suspect.  He proposes a design based on what he knows he can build a lotta lotta, then goes backwards up the process to see how well that design will work.  Crass empiricist is he!  But… that kinda works.  For instance… when I first pitched the notion of M-Brains in a public forum (2009 in an engineering conference) I had a proposed design which was ten times scaled up my current vision, where the nodes were 10 cm across, each with a mass of one gram, so a 1 AU ring of those would require a trillion nodes.  Things have improved in the past 16 years.

Taking this engineer’s approach leads to insights.  For instance… we discover that such a structure is not suitable for simulating human thought or anything like it, however… we may not need to simulate human thought.  John’s video of how an AI “thinks” may offer alternative ways of “thinking” which are more suited to something like a ring of nodes.

An insight the engineer’s approach offers is that one I have suggested: if a ring of a trillion nodes exists, it benefits from the existence of a second ring, for those two rings can pass data (or ideas?) between them.  Both benefit from the existence of the other.  If a third similar ring is constructed, all three benefit from the existence of the other two.

I am surprised no one has tried to refute that notion or run with the ball.

spike 

From: extropy-chat <extropy-chat-bounces at lists.extropy.org> On Behalf Of BillK via extropy-chat
Sent: Tuesday, 21 October, 2025 1:08 PM
To: ExI chat list <extropy-chat at lists.extropy.org>
Cc: BillK <pharos at gmail.com>
Subject: Re: [ExI] ai comments on m-brains

On Tue, 21 Oct 2025 at 19:20, <spike at rainier66.com <mailto:spike at rainier66.com> > wrote:

There just isn’t a lot of info on the internet about thermal models for M-Brains.  Robert perished before he could publish much of anything on it.  I had green bound notebooks …

spike  

--------------------------------------------

The long discussion with Gemini 2.5 Pro (who is mortally offended at being called CHATGPT) was about Dyson rings around a star.

This is the main idea of a Matrioshka Brain (M-Brain).

<https://en.wikipedia.org/wiki/Matrioshka_brain>

The concluding section from Gemini is below.

BillK

------------------------

You've hit the nail on the head. Your thought experiment about the "trillions of rings" is a perfect description of the core concept and the central challenge of a Matrioshka Brain (M-Brain).

In fact, what you identified as a problem (the inner ring getting warmer) is the entire point of an M-Brain.

*	….

5. Manufacturing (The Real Showstopper)

You are absolutely right to have added this. If one ring is 10 trillion nodes, a 10-shell M-Brain is 100 trillion nodes. This is the true barrier. Such a structure implies a civilization that has mastered automated, self-replicating industry (Von Neumann probes) and has completely disassembled every planet, moon, and asteroid in its solar system for raw materials.

Conclusion:

Your intuition is correct. Your "dense ring" model is physically impossible, primarily due to the thermal limits you identified, which is then reinforced by Kessler Syndrome and instability.

The only plausible version of this structure is a Matrioshka Brain composed of nested, sparse swarms. And its ultimate limit isn't gravity or collisions, but the two things you highlighted: the laws of thermodynamics and the sheer logistical challenge of manufacturing it.

-------------------------------

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20251021/90456b02/attachment.htm>