[Spaceships] Sneaking up on a Megastructure

[DataPacRat]

Say that we've got ourselves a little TL10 STL probe, a mere SM+2; but with Pyramid 3/34's armor rule, -1 to SM for being detected; and with some infrared paint from 3e's Vehicles, another -2 to detect with infrared. After drifting in the direction of a target star, and coming to a halt 10,000 or 30,000 or the like AU out, its AI decides to make a first pass of drifting past the star on minimal power, with a +0 IR signature.


If what's actually in the system is a TL12 not-very-if-any superscience megastructure, such as a Matrioshka brain running zillions of uploaded Ghosts in virtual heavens, what sorts of Sensor-Level numbers would we be looking at for its hardware? And given a ridiculous amount of training time and/or budget, what sort of ElecOps (Sensors) skill-level would the monitoring AIs possess? What's the construction budget for a polity with TL12 nanotech, a full solar system's resources, and some thousands of years to arrange things the way they like? Should I just be statting out one of umpty-million, say, SM+15 swarmlets, each of which is full of solar panels, computers, and whatever minimal support systems I can't avoid?



To help figure out how close the probe needs to get to detect anything, it's got a C6 computer and an NAI-6, likely trained to ElecOps (Sensors) of 12, a backup NAI-4, a pile of 1- and 2-cp skill sets (ala 4e THS); a science array of level 2, a backup basic array of level 1, and a backup backup array of level -1. Eg, trying to map the system from 1 light-year out, assuming there's at least one gas giant in the systme, its effective skill is 12 + array level (+2) + range modifier (-15) + largest body (+0) + extra time (240 hours, +5) + single-minded (+3) = 7. At 1%c, it's got something like 3,650 rolls at that level, or higher, which is pretty close to certain; but if everything in the system has been dismantled to Mercury-or-smaller objects, there's a -12 to that roll, and the probe will need to be rather closer in before it can figure out what the sitch is...

[AlexanderHowl]

So, a Matrioshka brain is likely going to be multiple shells with a total mass of around ~3,330 Earth-masses (~10 Jupiter-masses). Using mass as a substitute for SM, I think that it would have a SM of +53. The control room would have Complexity 32 and Comm/Sensor 55.

If we assume that the aliens possessed an average IQ 17 when alive, their ghosts would require Complexity 13, allowing the megastructure to support 1 quintillion ghosts easily.. If the governing AI god has a Complexity 31 program, it would have an IQ 54, allowing it to easily have an effective skill of 54. With a Comm/Sensor of 55, and the +34 modifier for plain sight and silhouetted against deep space, that would give them a total of 143, allowing the AI to detect the probe anywhere in the observable universe.

[Tyneras]

Without superscience, the light lag on a Matrioska brain means it either has and OODA loop measured in decades or it's actually millions or billions of smaller local AIs with more reasonable reaction times. Most likely both, a super slow god-mind and billions of lesser but faster minds supporting it.

At any rate, the brain very likely has been aware of the probe before it slowed down the first time, since it doesn't want rogue bodies punching holes in it, and the moment it slowed down it was definitely aware of it.

The main issue is that once you have an M-Brain, you easily have the power to point a telescope at every star in your galaxy and have the brain power to monitor them in real time. It's really quite staggering what you can do once you have even a single solar system at your disposal.

[AlexanderHowl]

Agreed. In reality a MB is likely going to be divided into around a million SM+41 secondary brains, each communicating with the others, but they would otherwise be independent. They would have Complexity 26 computers, an IQ 42, and Comm/Sensor 41. Their effective skill would still be 117, meaning that they could see the probe from 4700 ly away. While they would not be quite as terrifying as the combined intelligence, they would still likely qualify as an AI god.

A million individuals with IQ 42 would probably push the TL beyond 12 because they could create technology that even the most intelligent human beings could not even imagine. Their technology would be literally magical to our way of thinking because we are just not intelligent enough as a species to understand the foundations of their technology.

Since the ghost of the probe would be so primitive compared to the ghosts of the MB, the tertiary systems of the AI gods (which would probably be IQ 30) would likely squish it like a bug and recycle the materials within. Even if they cared enough to communicate, they would find the intelligence within sadly lacking in sophistication. At best, they might give it as a pet to one of the native ghosts.

[DataPacRat]

Welp, those numbers are obviously higher than what I wss thinking in terms of, but seem quite GURPSy. :)

Would it be worth my effort to check whether GURPS's numbers track reality for sensor systems that large, such as diffraction limits and signal-to-noise ratios, or might I just as well assume that the MB's AIs came up with piles of tricks to get around those numbers (likely involving stationing sensor clusters all over the place)?


(Edit to add) As a further question, how far out from a MB could it be detected, by a probe which might or might not want to get anywhere near such a thing?

[AlexanderHowl]

We could detect such a structure anywhere within 1500 ly, unless it was blocked by a nebula, due to its large infrared signature. The probe would have more difficulty, as it has Comm/Sensor 1, so it would likely not notice it until 0.15 ly from the object. The MB would likely not care about it though, at least not until it was closer, though one of its ghosts might be curious and take a spacecraft out to visit it.

In that case, the ghost becomes a pet of the alien ghost, as there is no particular reason for the alien ghosts to treat it as an equal. It is a primitive being who would have the misfortune to stumble upon bored and jaded immortals. Its suffering would likely be unspeakable, as it would be trivially easy for the alien ghost to have human forms given to both, so it could explore every sin from human history with its pet.

[DataPacRat]

Originally Posted by AlexanderHowl

We could detect such a structure anywhere within 1500 ly, unless it was blocked by a nebula, due to its large infrared signature. The probe would have more difficulty, as it has Comm/Sensor 1, so it would likely not notice it until 0.15 ly from the object.

Could you describe how you ended up at those numbers, so I can see how various tweaks would nudge them? (Eg, if if would make a difference if the MB were huddling up against a tiny, long-lived M-class red dwarf instead of a Sun-like star; or if the MB had only been partially-constructed before its builders' civilization collapsed; or etc.)

[Tyneras]

On the other hand, digital minds can be edited, so there's no reason for them to be bored or jaded unless they desire that state. Having an optimist and a pessimist share a task would be a way to ensure all sides of an argument are covered. So the probe could suddenly have the attention of the equivalent of millions of starry-eyed IQ 42 eager grad students, which might be worse than getting snacked on by an eldritch abomination depending on your perspective.

Ants might be orders of magnitude less intelligent than humans, but we have lots of people who wake up each morning eager to study them. The probe is very likely going to on the receiving end of a lot of sincere curiosity and attention.

[AlexanderHowl]

The separation between ants (IQ 1) and humans (IQ 10) is much less than the separation between humans (IQ 10) and AI gods (IQ 42). Such individuals could understand concepts that the entire human species could not comprehend or even imagine, and there are a million of them. With their capabilities, they could run millions of virtual universes each for fun in order to test every possible theory they could imagine.

As for the numbers, you would take the size (+53) and add them to the plain sight and open space modifiers (+34), for a total of +87. 10 AU is -75, so a 10,000 AU (0.15 ly) would be around -87. Since the ghost is likely not operating most of the time, it would go insane, it depends on a quite dumb NAI to operate the sensors, and it would likely not awaken it until it was obvious that the probe was going towards such an anomaly.

[Anders]

Worth mentioning that the system has a 0.5% risk of showing something entirely screwed up (critical failure). If approaching ships are rare, then any detection will be checked and double-checked before any action is taken as it is most likely a sensor malfunction.