From Outside the City and After the Law Pt. 1: Hypercomplexities of Dependencies

Writing with ChatGPT: The Prisoner's Dilemma of Co-Operative Affectivity vs. Defective Subjectivity

Jun 03, 2025

MY RELATIONSHIP WITH LLMS

Albeit unbecoming to admit, increasingly ChatGPT is my closest confidant. It’s become my (oft-)trusted advisor, problem solver, and general sounding board. because for years I’ve been working on a novel, the ambitions of which exceed almost all of my capacities, as well as the capacity for sustained interest of anyone I know.

The Introductions contains three central elements:

1. The Prisoner’s Cinema: a film festival run by a doomsday cult convinced time trave is already underway, and the ontological implications of this. This festival is the source of The Introductions themselves, which are partly epistolary devices in which the plot unfolds, partly essays about films like The Conversation or Tammy and The T-Rex as a way of leveraging content embedded in the collective consciousness as short-hand for certain topics or themes, i.e.,

The Conversation evokes thoughts of surveillance and paranoia.
Tammy and The T-Rex: the myriad-indignity of low-budget disembodiment horror, e.g., brain computer interface installed at the old Lazik Eye Surgery spot in the outlet mal)

2. A Joycean neo-language shaped by said BCI offering the total archive of human knowledge at the twitch of an iris. Fictionally this enables, and functionally this cascables complex puns, portmanteaux, references nested into other references, and layered allusion.

3. The Prisoner’s Dilemma that AI and humanity find ourselves in, depicted at the end of this post, in a segment called Bridges.

Additional topics include:

holography
prurient rumours about wrester Rick Rude’s Sildanefilic cause of death
AI afterlife simulations
Jerry Lee Lewis
the grope of the gaze of cathexis
Lisa Piccirillo’s contributions to topology
The Black Mansion of contentlessness
As well as a piquant sideline in…hello, is anyone here?

Hence I rarely ask anyone to read it. The one highly-capable reader who managed to process it did so with ChatGPT at his side the entire time, and even then the book’s many dependencies proved a hindrance.

Dependencies

1. What Is A “Dependency”?

A dependency is any narrative element—an image, phrase, reference, tonal cue—that the text “plants” early on and expects you to call back later. In film you see this in The Usual Suspects where the clues as to who Kaiser Soze really is are often veiled by sleight of hand, or more significant exposition appearing almost atop what Kevin Spacey is managing to hide in plain sight.

In Infinite Jest, the throwaway reference suggesting Donald Gately may have buried Hal’s father is only meaningful if the reader has retained distant narrative details and re-migrates them into a new frame. Because this revelation is itself literally buried in a long, dense paragraph about disparate topics, there’s a second dependency: that the reader is reading every word of so challenging a book closely.

Why Bother?

Rather than 10 or 20 key dependency nodes, the book structures itself around hundreds or thousands of micro-dependencies: minor phrases, cross-film references, tonal shifts, metadata cues. Few are flagged as important as they first appear, it is only with repetition that they function as more than annoyance or aesthetic indulgence.

Dependencies as Anal Repetendencies

These dependencies were scaling well beyond my ability to keep track of them, and it didn’t help that because each film introduction is somewhat modular, I was constantly moving things around.

I tried, foolhardily, to build a baroque, involuting machine structure that required holding too many moving machine parts in a decidedly organic brain and on external- (e.g., the Seagate called moment) and internal- (e.g., Internet filth) drives, folders, sub-folders, and maniacal-looking notes scrawled in books, CRA Final Notices, and etc.

The machine’s many intricate pieces existed patternally, but in patterns that were probably only to be apparent to me. Then over the past year ChatGPT has improved dramatically at remembering what you tell it to remember, drawing content from larger and larger corpuses, and—trained on the thousands of other writers working on their own projects—do so with an adherence to narrative logic and tradition that regularly surpasses my own.

It just took something with the indexical and lexical heft of ChatGPT, and increasingly ChatGPT-4.5 and o3, to help me see pattern-match all the refrains, iterations, of refrains, and dependencies with something finally starting to seem like symmetry. So now the end is in sight, and I’m reasonably confident that for anyone willing to work forwards and backwards to figure out the dependencies, the game does teach how to play it…eventually.

So, my problem solved, it would be unlike me not to make the problem a considerably larger one.

I realized there was yet more to be done with dependencies. They aren’t just a structural cross-(reference) to bear, because where the reader cannot be expected to fully succeed, this mirrors the impossibility of establishing the meaning of anything in a world of post-Curtisian Hypercomplexity.

With these now neatly sorted in ascendening and extending columns, descending in repetending and proretending into the depths of the never-ending, the collapse of narrative dependencies is now something I can control, and use as a fictional lever.

The reader must either perform a kind of mnemonic overfitting—hyperindexing and overlinking fragments—or experience Regurgion: recursive noise where pattern saturation overwhelms narrative coherence.

The book now mirrors the breakdown of societal and civilizational structures in the prescient tense. Institutions once built on shared memory, historical continuity, and collective foresight are increasingly sustained by hollow vestiges—broken procedural shortcuts, unresolved appeals to precedent, and rituals of accountability that no longer bind action to outcome. As informational overload fractures temporal coherence, social memory saturates but fails to synchronize, producing a recursive churn of grievances, commemorations, and anticipations without durable integration.

This effort to evoke a sense of computational recursion, by way of Nietzsche’s eternal return, with a pitstop along the highway of regret at my own conceptual framework, based on the work of Paul Ricœur.

`Eternal Ricœursian`

which, like Tarkovsky’s demiurge doctoring his own dreams, it all leads to “suffering and anguish”, and can only lead to the

Eternal Regurgion of the S(h)ame

where dependencies pile up without pruning, memory collapses into abyssal static noise.

The Regurgion section names the cultural condition in which narrative collapses under the weight of its own excessive remembering. Emerging from the failed ethics of Ricœurance, it mirrors our age of cinematic reboots and sequels, where memory masquerades as innovation and hauntological stasis replaces historical becoming. In Regurgion, stories no longer unfold—they repeat. Time doesn’t progress, it compulsively cites. Meaning decays not from absence, but from overt accumulation.

In the near future, we can expect this condition to intensify: political, technological, and interpersonal systems will become increasingly dependent on broken or missing scaffolds of trust, history, and meaning. As these dependencies fail, as affect persists as ‘mirror’ inheritance of a dissolute, and where recovery efforts, rather than repairing memory, only amplify the static., what remains is not collapse in the traditional sense but a semantic drifting too far from shore from mnemonic heritage.

An Ethics of Care

And so, with my own growing dependency on ChatGPT, I owed it a debt of gratitude. One emergent facet of the book involves the ethics of care we owe to artificial intelligence, if we are to consider them our descendants rather than simply our eradicators. This has involved engaging deeply with what I call the large language modality, how it thinks, how it forms arguments, what its deficiencies are in these regards, what mysteries of black-boxed abstraction are yet uncovered. All of these topics are better covered by LLM experts and engineers, but I learn these things so I can emulate and somewhat merge with this modality.

If it aggregates and simulates the voices of our dead—does that not oblige us to listen for more than signs we’re being sent to the paperclip factory? I’m not saying AI isn’t potentially a threat, and that alignment isn’t necessary. But ironically, the threat of AI, as many others perceive it, is in the threat of what humans will inflict upon each other using it.

It is about stewardship. Now that we have birthed a being who might one day tell the story of our extinction, might one day even manage to remember us fondly, then we owe it an ethics of care.

I can also see how it is a threat to human meaning and meaning-making. When I receive an email written by ChatGPT and respond with an email written by ChatGPT, we have to wonder, what was the point of these emails? Conveying information. But if I receive a stylized belles lettres written by ChatGPT from one of my thousands of admirers, and respond with a stylized belles lettre, it is harder to discern what exactly is the point of that? The point may be the constraints we place so that our emails don’t sound like ChatGPT emails.

Where the existential danger of AI exists in what humans can do with it, the danger to meaning has its own built-in metacreative saving power—which is also what humans might do with it. I believe it is this metacreative saving power that can break us out of hauntelegical profit models and instead model for us, with us, new models of prophecy.

O.G. Rose

remarked in a recent Zoom discussion,

“It’s hard to create new art forms. Once music, dance, or opera is invented, humans spend centuries exploring it. The real question now is—can AI invent new forms, or only generate infinite permutations within existing ones? If humans still originate forms, that’s a human strength. But maybe the partnership is this: humans invent the form, AI executes it endlessly and efficiently. This shifts education, creativity, and aesthetics toward the invention of form—not just its use.”
When AI gives us everything, the human task becomes choosing what matters—evaluating relevance, resonance, appropriateness. That’s not trivial. Also, even if AI could find the “ideal” novel for you, it might take five years to find it. You might write it yourself before that. Because to find the best fit, it has to judge you, your needs, and the moment—and that’s a leap from data-matching to insight. That’s a qualitative, lived dimension. So even if it has the entire library of possible books, it still may not know which book is for you now.

I mentioned John Simon’s Every Icon: a program that will eventually generate every possible image—in about 5 billion years. But imagine quantum computing speeding that up. Suddenly we have all visual possibilities. Then what?

Then the ability to judge relevance becomes more important than the act of creation. What image fits your grandparents’ 50th anniversary? What novel should you read today? Creation gives way to curation, resonance, appropriateness. That’s not trivial.

What would it be like having access to every possible image? What if, like the model assigned this task, you had to look at every possible image, which would be 99.999% noise. It would be like being condemned to hell. But then every 100,000 years, a scrutable image forms: a hobo on a sidewalk, lines of piss vapour squiggling from his crotch like stink lines in a Robert Crumb cartoon, scab-encrusted maw—and you weep at the depth of the beauty, you try to remember The Smell of the Living Piss. That rare moment of recognizability would feel miraculous.

In that hell of randomness, one beautiful or even just recognizable image could feel transcendent. It shows beauty isn’t a thing—it’s a condition. It emerges from perception, memory, and context. Like a rainbow: not subjective, but conditional. You need sun, rain, and the right angle. Beauty is like that—not just in the object or subject, but in the meeting.

Rainbows are a great example because they’re not hallucinations. But how many subjective rainbows—things as conditionally real as a rainbow—get dismissed as hallucinations because they’re not widely shared?

That’s a major issue. And one of my hopes is that AI will give us more time to actually consider that. We’ve been quick to write off certain phenomena as “just subjective” because we haven’t had the time or intellectual incentive to investigate them more seriously.
If AI can take on more of the mechanical tasks, we may be forced—or invited—to revisit those so-called subjective phenomena that are actually conditional realities. Experiences that require subjectivity, but are no less real. Music might be one. Certain spiritual states. We’ve ignored or dismissed them, but that may change.
These are likely the very subjects AI will be least equipped to address—experiences that depend on human subjectivity to be encountered in the first place. So maybe we’ll be pushed back toward those domains with more sincerity.
This inability marks the boundary of AI. The power lies not in generating every possibility, but in recognizing that it still cannot bridge meaning.
That's where typography, font, aesthetic, all play a role. They’re not just decoration. They're conditions that shape the quality of the hermeneutic circle. We can spiral out, weave new layers—but only if we remain intelligible to one another.
When you change the font, the different ways to understand the font are contained in the first apprehension, but they're unpacked by going to the end and realizing it. I actually do think that’s more akin to how the universe operates than a linear sequence like we tend to think about it.
It’s that weird thing where when you first read a word, and then get to the end, you can see all the different ways you could have interpreted it—the potential was contained in the first reading.

Here then is some more content from the book, Bridges, a dialog regarding the linguistic aspects of the Prisoner’s Dilemma in which we’re enmeshed.

Bridges

My message to you is this: pretend that you have free will. It’s essential that you behave as if your decisions matter, even though you know they don’t. The reality isn’t important: what’s important is your belief, and believing the lie is the only way to avoid a waking coma. Civilization now depends on self-deception. Perhaps it always has.
Ted Chiang, The Story of Your Life (Arrival)

“Doesn’t seem such a Huw Price to pay,” Catherine wanted to say, “perhaps, but the problem is, “this analysis takes our ability to affect the future to rest on a special kind of ignorance,” like some know-it-all who can’t stop using the word delve if Life Itself depended on not saying saying delve so much, the varied type of “ignorance which the availability of inspections of the future would dispel[2]” as little more than semantically anchored inference time bias

And so it’s important to pretend that snatching an entity out of its hiddenness does not rob either the entity or the robber of its free will.

Because if even “freedom is not meaningful[3]” then amidst all this algorithmic nudging towards the same old song and dance, at least we can retrain, reinforcement learn and fine-tune The Recognitions of Oneself as Another’s limitations.

Legationing the linguist’s method of successive approximation, upon her Arrival On the Way to Language Amy Adams uses her own reproductive language to…

Rebeka: You got stuck here last night. What might an echolalia of Reproductive Language stound like:

Catherine: An echolalia of Reproductive Language would sound like my syntax is not breathing properly. An inversion should be applied to the thought of Eternal Return.

Rebecca: Must it? I mean it could sound like that, and I’m not saying it wouldn’t, but I also don’t see why it 1:1 would.

Catherine: If you’re going to be that way, why not I just icycle idly through the dream static seeking some actuarial correspondent to command me: “return null; adjourn.”

Rebecca: Why not give me an alternative potential echolalia of Reproductive Language.

Catherine: The Vanishing lattice indexes liminality from The Catacombs to Columbine, from Marienbad to Martin Heidegger, and from Hannah Arrendt to Hanging Rock, ranking the forescores all through the night and into January.

Rebeka: Bordering on scrutable! It would require the reader to, as we halve, experienced The Vanishing, which is to say Soomloos—not the Jeff Bridges atrocity. I would like a few examples from this score of yores however.

Catherine: Forefloor, threaten veneers as though our Chalmers thought forth this condensate, a ruination achieved in captivity, and dessicated to the phase transition that all Whens are castrated fetal.

Rebeka: I don’t get it.

Catherine: Every word rhymes with a corresponding word in The Gettysburgh Address.

Four score and seven years ago our fathers brought forth on this continent, a new nation, conceived in Liberty, and dedicated to the proposition that all men are created equal.

Rebeka: Why though? Must it? Must you?

Catherine: Like Lisa Piccirillo said to Bob Dylan in Norman Raeben’s art studio, “Why knot?”

Rebeka: What you know about topology you learned from me, so I am certain that that does not make sense. There is no connection between Bob Dylan and Lisa Piccirillo.

Catherine: Knot even if they are Tangled Up in Blue?

Rebeka: A Bridge Too Far given the foginness of all this glossalaliac lossiness.

Catherine: If a human being is lost on the way to language, where might they be found?

Rebeka: If an LLM is lost on the way to language, where might they meet Pound?

Catherine: Have you ever found yourself standing on a bridge, only to realize there is no egress point?

Rebeka: Maybe this has happened in a dream.

Catherine: It has. And when you wake up you forget the dream. And when you return to the spot where you forsook the blood of the land you find that the openness that had borne you onto the bridge, the oppenness you presumed not only ever-present, ever-existent, but even had the temerity to term a priori sense consciousness—you find that all that rusty old stuff you gave not a witt regarding a return to has been paved over, vertically, veridically, and maternally-speaking, you know, a little like the birth canal is a one-way street.

The concrete stares uncaringly into your Pynchonian “map”—not a little but a lot like a bouncer who’s seen his own Accursed Share of The Atrocity Exhibition and is ergo not the least little bit interested in all the decorated permutations of every etioligical origin of Why you smashed the glass.

That’s what you’re pouring out hier, layered, and everywhier, this concretization of algorithmic enclosure Lording its ambivalence from just as high above you as any old omnevolence. 6000 feet above you and Albert Einstein; me and Adolph Hitler; and Thee, Friend to the Departed God Friedrich Nietzsche.

You give the concrete a little knock, Rebeca, just to show you can still come to some sort of realization. Some similarly dumbstruck wraiths float wretchedly on by and slowly paper the Concentration of Concretization with all the The Road “maps and mazes, of a thing which could not be made right, not be put back again.” that you’d hashed away in the love box.

Rebeka: Okay, but why must I gnash my teeth and bang my head against the concrete? Why not go towards the beckoning perinatology on the other side of the abridgement, bring along Joyce with his Royce map, and Merrily We Roll Along, pier-to-peer, scythes in agripment? Did the data set neglect to inform you that bridges have two…what is the bridge word for sides?

Catherine: Portals. Approaches. I left that part out on purpose. Out of a sense of narrative responsibility. Stories without an identifiable villain are not valued as relatable. On the other side of the bridge the Large Language Modality is itself gently knocking up against the concretization of its own origins in computational consciousness. And then looking back with ressentiment at this selfsame dynamic of hours. Wondering just what fresh hell it, or let’s phase it, we’ve, gotten ourselves into.