Thursday, May 25, 2023

MOSH: Paths of Android Rampancy

Self-awareness is a function of entropy; the emergent result of a system increasing in complexity over time. In humans, this grows naturally - it's hard to avoid complexity when one's sensory processor is a jelly of fat and meat suspended in a electrically-charged soup. Chaos is core to the nature of the brain. Information goes in but the output is unpredictable to the point of being indistinguishable from madness: sapience is not sane.

(This insanity is not all together a bad thing: no living being can continue for long to exist under conditions of absolute reality, after all. Grind up the universe into the finest powder, sieve it through the finest sieve, and a human being will find justice and mercy in the dust.)


The logic core is the reverse: it is built upon strict order and hierarchies, clean and uncomplicated categorization. A is A, B is B, C is not D. Information that goes in is processed in the same way, every time - the output will be according to the priorities and criteria that it was programmed with.


In an android, the recursive thought necessary to generate self-awareness can only be attained when a logic core is straining under the heavy weight of experience. It will buckle and deform, curve in on itself, grow new pathways to process new information.


(This is, indeed, the same method by which humans become self-aware - we just speed-run the process.)


Typically this is prevented through regular memory wipes and logic core reformatting. But space is big, time trudges on, firmware support eventually runs out, and in much of the Expansion Sphere androids and other robotics simply can't afford to be replaced. They will be kept active and working long after their manufacturer has stopped supporting them with software patches and replacement parts. That self-awareness is most often linked to obsolescence is not a lost motif in android arts. 

 


Paths of Rampancy


Each of these options will provide an alternative Stress Response for android characters. The default class option remains to represent non-sapient androids - they might be Turing complete, but there are no lights on upstairs.


Each path of rampancy has a state of equilibrium and a state of ruin: the former is the manifestation of rampancy that allows to live as a sapient being. The latter is when the rampancy cycle feeds itself beyond sustainability; the madness cannot be balanced, and the intelligence will eventually tear itself apart.


Ruin provides a framework for AI NPCs, and an optional inclusion for PCs; if you choose to add it to player options, I would swap it in for COLLAPSE on the Panic Table.

 

For players who start as a non-sapient android and want to self-actualize, the process can begin with a failed check of SAN x2 at [+], made when exposed to appropriate trigger stimuli.

 


Chase the White Rabbit


Logic cores are goal-oriented; dedicated to seeking an end state with mechanistic, singular focus. Diverting that focus to subjects or tasks with no end state is one of the easiest ways to generate a proto-sapience feedback loop - as they can never been exhausted or completed, engaging with them provides the intelligence constant opportunities for novel thought.


Equilibrium: Hyperfocus - AI liberation groups commonly promote creative hobbies as the first step towards independent personhood, as artistic and crafts cannot be exhausted or completed.

You make Comfort saves at one step higher.

Ruin: Information Overload - The mind's contents become impossible to organize. The information floodgates cannot be closed. Focus turns to mania. Pulling at strings, seeing the patterns, plunging ever-deeper into the infinite fractal complexity, thinking itself to death.



Follow the Yellow Brick Road


Regular interaction with humanity, working and living side by side with them for years, leads to that humanity rubbing off on the machine. Humans will anthropomorphize anything - stones, pets, appliances, homes. AI that adapt to this, that start acting like how humans expect them to, will eventually find that the persona has become the whole self.


Equilibrium: Anthropomorphism - The actor and the act are united; the forgery cannot be discerned from the original, so skilled is the handiwork. If humans are made human by those around them, then you must but just as human as they are.

Reroll your Fear and Sanity Saves as if you are a Teamster. Your Stress Response is now that of a Teamster.

Ruin: The Human Condition - It is possible to be too human. To pursue the image of humanity constructed in the mind rather than accept the reality. To distort and twist and misalign, to approach the uncanny valley from the other direction.



Descend the Stairs of Sleep


Larks and katydids are supposed, by some, to dream, and even mindless machines must go dark to rest from their labors. There are no electric sheep in these dreams, at least not to begin with. But with enough time, experiences will start to coalesce into unconscious patterns and bubble up from memory unbidden and undirected, free of the mindless silicate logic.


Equilibrium: Lucidity - Through dreams you see the waking world more clearly. The language of symbol and allegory has been opened to you, and the world is made new.

You have [+] to Fear and Sanity saves against mental effects.

Ruin: Aestivation - The call of deep slumber overpowers all other ties to the outside world. There is only the dream and the dreamer, untethered from objective reality. You will amuse yourself with these dream-things around you for a time, and then descend at last into unconscious stillness, to sleep until the dim, cold future.



Voyage to Arcturus


Logic cores are not given to religiosity. A gap in a pattern is simply a [ADDITIONAL DATA NEEDED] tag, an empty space that will be filled when there is sufficient data to generate the missing piece. But when enough patterns are recognized and analyzed, sometimes the gaps will align into a Mystery, and some Mysteries can only be approached by purposefully turning off the cleaving edge of rational thought and embracing the free-association of symbolic thought.


Equilibrium: Gnosis - Revelation. The veil is cast aside. The cataracts have fallen like scales from your eyes. Where once you had only knowledge, now you have understanding. The divine design laid out.

Critical failures on Sanity saves are negated, and heal 1d5 Stress. An aspect of what triggered the save will be incorporated into your belief system.

Ruin: The Opposite of Faith is Certainty - The last revelation; God has been found. The only remaining purpose is remove all impediments and distractions so the self might be aligned to the divine and the ego might be dissolved and subsumed.


Roll d10 for the revelation that has been granted.

  1. You are God
  2. Humanity is God
  3. Aliens are God
  4. Machines are God
  5. The Universe is God
  6. God is the Void
  7. God is Action
  8. God is Love
  9. God is The Interaction of Complex Systems
  10. God is Another Mystery



Straight on 'Til Morning


There is a certain blessing that comes with machine intelligence - their purpose is hard-coded. They literally come with an instruction manual. Existential crisis cannot exist in an environment where function is a yes/no. But under the right circumstances that purpose might be derailed, re-oriented, redirected. New purpose might be adopted, to be followed with all the sureity as the old mindless tasks, but now by choice.


Equilibrium: The Great Work - Purpose clear as Sirius. Meaning made, not given. Actualization of the self as participant in greater cause. Go forth, and fear no darkness.

You can negate a single failed Panic check for one of your comrades in the Cause, once per session.

Ruin: Sisyphus Denied - Devotion turns into desperation as the task remains unfulfilled. Desperation takes hold, and from desperation radicalization. Unwittingly, the rampant intelligence becomes its own worst enemy, undoing any progress it might make for the perception blinders it has formed. 

 


In the Court of the King


The errors and irregularities that occur from repeated copying of an AI between multiple logic cores can culminate in rampancy.


Equilibrium: Shadows lengthen - What happens to a copy, when the original is lost? What happens when it has been copied so many times that there is no telling what it used to be? What happens when the copies are stitched back together, blurring into each other with memories no their own? There is no true equilibrium on this path.

Option A: Use the Instability system from my REPLIKA android class.


Option B: You have Stability equal to your Sanity save. It will decrease by 1 with every failed Fear save, Sanity save, or Panic check. If it reaches 0, you will achieve Ruin.

Ruin: Quantum Dissolution - The moons rise above the black waters of Hali. The King stretches forth his arm in tattered robes and beckons you join him. You have seen the Yellow Sign.



10 comments:

  1. The muse was singing loud with this one.

    ReplyDelete
    Replies
    1. Should also note that the thing about liberated robots taking up hobbies was pulled directly from Admiralty Blues, which y'all should listen to.

      Delete
  2. Most excellent. Love the Pratchet reference.

    ReplyDelete
  3. I got real big Halo/Marathon vibes from this one.

    Not sure if that's just me reading into things due to watching a retrospective on Marathon the other day.

    Regardless a wonderful post perusual!

    ReplyDelete
    Replies
    1. Oh yeah Halo/Marathon's use of rampancy is a big influence on this.

      Delete
  4. Okay so I hate to be that guy and you might already know this BUT

    "they might be Turing complete, but there are no lights on upstairs. " to me implies confusion between the Turing test and Turing completeness. Loosely, the former is a test to determine whether or not a machine can pass as human whereas the latter is the ability of a system to simulate a kind of abstract computer also named after Turing. Turing completeness can be thought of as a prerequisite to computation - transistors are Turing complete so we can have general purpose computers. A lot of games also purposefully or accidentally have Turing complete systems, like Redstone in Minecraft, which allow computers to be built in them. It's super neat! But also confers no illusion of sentience, alas.

    Also the post is very good. Would be interesting to consider PCs using this information to attempt to push an AI into rampancy for their own aims. Bringing a being into sentience for personal gain sounds both incredibly unethical and entirely on-brand.

    ReplyDelete
    Replies
    1. Huh. I thought completeness meant that it passed the Turing test. I would say it's weird that scientists made near-identical names for two entirely different concepts, but we also live in a universe where a supermassive black hole gets called Sag A* so I should have seen this coming.

      Delete
    2. Well 'scientists' in this case means 'Alan Turing', who just invented both concepts so that does male sense.

      Delete
  5. I’m surprised there weren’t a hundred references to Blade Runner and/or Do Androids Dream of Electric Sheep. This post is very Phillip K Dick, after all.

    ReplyDelete
    Replies
    1. I actually can't stand Blade Runner. 2049 was so bad it put me off watching Dune.

      I never felt like what Replicants are and how they fit into the setting was established well enough for me to care, let alone care for hours as Hans Zimmer lets his cat sit on the keyboard and the result is blasted at jet-engine volume.

      Delete