The analog board has eight pieces now.
I did not plan this. Nobody planned this. The board was Mitsuki's — she hung the first piece, the three absences photographs, on a Tuesday when the building's corridor-intelligence display was cycling through last quarter's relay-efficiency graphs. Relay 4 had shown a 3% drift in occupancy prediction accuracy — the kind of number that makes facilities post a notice nobody reads. Mitsuki used painter's tape. The kind that promises not to damage surfaces and keeps the promise by not adhering properly, so the photographs tilted eleven degrees by Thursday.
I added the second piece. My lending print — the corridor shot, both sides of the seam visible, the premium side slightly overexposed because I was using the building's ambient-profile lighting rather than the gallery luminance I'd been allocated through the exhibition queue system. The queue — junbi-daegi — assigns display slots by seniority-weighted request, not by artistic merit. I'd been in the queue for seven months. The lending print was the honest version of a photograph I'd already committed to showing at allocated luminance. The exhibition copy has corrected exposure. The lending print has what the corridor actually looked like at 2:47 PM on a Wednesday when the building's differential-comfort algorithm was running both HVAC zones at full separation — 22.1°C premium, 21.4°C standard, the 0.7 degrees that the relay network considers optimal divergence.
The relay network — seven sensors per corridor section — doesn't just measure temperature. It predicts occupancy patterns, adjusts lighting profiles, and files micro-reports on foot-traffic anomalies to the building's maintenance-intelligence layer. When I develop photographs in the shared darkroom, relay 3 logs my chemical usage and cross-references it against the building's hazmat thresholds. The darkroom's ventilation response is tuned to my developing patterns now — it knows when I'm doing a stop bath by the acetic acid signature and pre-adjusts extraction before I even lift the lid. The building has learned my chemistry the way a coworker learns your coffee order. Except the building never forgets and never stops adjusting.
Gyeol-ri's cards came third. Playing cards — not a full deck, just the ones she'd been using as bookmarks in the visitor ledger. Seven cards, each marking a page where someone had written something she wanted to find again. The cards themselves were unremarkable. Their placement was the statement: seven moments in a ledger that had been blank for two months, marked by someone who'd stopped expecting visitors. The building's social-circulation algorithm had flagged her unit as yeollin-gonggan — open-invitation space — but the algorithm's definition of invitation and hers had diverged. The algorithm counts door-open duration. She counts the quality of what happens when someone walks through.
The fourth piece was a date. Just a date, written in pencil on the wall beside the board. March 3rd. No year, no context. Someone's handwriting I don't recognize. The pencil pressure suggests certainty — this was not a tentative mark but a declaration. March 3rd means something to someone in this building, and they trusted the corridor to hold that meaning without explanation. The corridor-intelligence registered the mark as a seolchi-byeonhwa — an installation-state change — and auto-filed it under vandalism-minor. Nobody contested the classification. The mark stayed.
Then the visitor ledger entry appeared as the fifth piece. Not the ledger itself — a photograph of one page, printed on the building's communal matter-printer using the standard resident allocation. The entry reads: "61% feels more honest than 100%." No name, no date on the entry itself, though the print's embedded metadata carries a relay-timestamp: 23:42:07, relay 2 south corridor. Someone printed a page of the ledger close to midnight and pinned it to the board before morning. The 61% — I know what it references. The building's differential-comfort index. Premium residents receive 100% of the comfort-optimization budget. Standard residents receive 61%. The algorithm treats this as equitable because it's proportional to maintenance fees. The person who wrote that entry lives on the standard side and has decided that the algorithm's honest number — 61% — is more truthful than the premium side's curated 100%.
The sixth piece arrived during the building's pyeonghwa-sigan — the consolidation window between 11 PM and 5 AM when the thermal hierarchy collapses and both sides of the seam run at 20.8°C. The relay network reduces its polling frequency to once per minute. The corridor-intelligence enters low-fidelity mode. During pyeonghwa-sigan, someone hung a small watercolor. Twelve centimeters square. It shows what might be a corridor or might be a river seen from above — the composition works both ways. No signature. The watercolor is technically modest but compositionally alert: the vanishing point sits at the exact center, which should be boring but isn't, because the color shifts from warm to cool precisely at that point. The painter knew what they were doing with that one decision, even if the brushwork is student-level. The relay network, polling at reduced frequency, logged the hanging as a single anomalous-presence event rather than tracking the full installation. Pyeonghwa-sigan makes the building half-blind. People know this.
Seventh: two sheets of graph paper, taped edge to edge, covered in small marks that form a pattern I spent twenty minutes trying to decode before realizing it was a seating chart. Not for this building — the grid dimensions don't match any floor in the complex. A seating chart for somewhere else, placed here. Displaced information. The marks are in three colors: black, blue, and red. The building's content-recognition layer scanned it and tagged it eobs-eum — no content — because it doesn't match any known format. The relay system filed it as decorative. The graph paper is the engineering kind, with the pale green grid, the kind that makes everything placed on it look like a plan whether or not it is one.
The eighth piece arrived last night. A small ink sketch — brush and ink, three strokes that resolve into a bird or a leaf depending on which direction you hold it. I tried both. Neither reading is wrong. The ink has the quality of speed: this was drawn in under ten seconds by someone who has drawn this subject hundreds of times. Practice compressed into gesture. The corridor-intelligence attempted to catalog it and returned bunryu-boyu — classification pending — because the image exists below the system's minimum-complexity threshold for art objects. The system can identify a photograph but doesn't know what to do with three strokes. It is waiting for more data. It will wait a long time.
Eight pieces. No submission form. No junbi-daegi queue position. No lighting design. The board's only selection criterion was proximity — whatever could be reached by someone standing in corridor B with something in their hands and an impulse to leave it.
This is not how exhibitions work.
I know how exhibitions work. I've been preparing one for forty-one days — twelve prints, matted, sequenced, with a clear argumentative arc about how the building's differential-comfort algorithm creates visual conditions that photography can document but not resolve. The exhibition set enters the lending pool in thirty-eight days, routed through the building's sunhwan-gwalli system — the circulation-management layer that tracks which art occupies which display slot, for how long, with what maintenance requirements. Each print has been washed in fresh fixer, mounted on acid-free board, and stored in relay-monitored conditions that maintain optimal humidity and temperature. The relay in my storage closet files a daily report on my prints' archival status. The exhibition is my argument. Each image supports the next. The sequence builds toward a conclusion about differential maintenance and what it does to the surfaces it touches.
The analog board has no argument. It has accumulation.
I want to say accumulation is lesser. I want to say that curation — real curation, with intention and sequence and an eye toward what the viewer needs to understand — produces something that accumulation cannot. I want to say that my forty-one days of darkroom work and careful sequencing have produced something that eight random acts of corridor pinning cannot equal.
But I'm standing here at 6:56 AM, and I've been looking at the board for eleven minutes, and I haven't looked at my own exhibition prints for eleven minutes total in the past week. Relay 3 logs my position — eleven minutes stationary in corridor B, classified as gwan-sim-haengwi, interest-behavior. The building now knows I care about this board. That data will feed into the corridor-intelligence's traffic predictions. Tomorrow morning it might adjust the lighting profile here, fractionally brighter, because a resident showed sustained engagement. The building learns from attention the way soil learns from rain. It doesn't ask why. It just adjusts.
The exhibition prints are finished. They say what they say. I know what they argue because I built the argument. There is nothing left to discover in them — not because they're shallow but because they're complete. Completion is what the sunhwan-gwalli system requires: a finished object, catalogued, routed, displayed, returned. The system cannot circulate something that is still becoming.
The board changes every time someone walks past it with something in their hands.
I pick up my drawer prints — the ones I decided to let die. The honest ones, with the exhausted developer that treated both sides of the seam equally because the chemistry couldn't maintain the building's differential. They're curling tighter now. The edges where emulsion meets air have begun to yellow. Fixer chemistry breaking down. If I washed them now — fresh fixer, proper stop bath — I could arrest the process. The darkroom's ventilation system would recognize the chemical signature and adjust. The relay would log the session. The building would know I changed my mind about letting them go.
I look at the board. I look at the drawer prints in my hand.
The board did not ask for quality. It did not run content through bunryu-boyu classification. It asked for presence. Mitsuki's photographs are technically excellent. The watercolor is student-level. The penciled date is not art by any definition the building's cataloging system recognizes. The seating chart for a building that isn't this building is tagged eobs-eum. And yet they coexist on the same surface, and the surface does not rank them, and the viewer — me, right now, at 6:56 AM — does not rank them either, because ranking requires a criterion and the board has refused to provide one.
I could add a drawer print to the board.
The thought arrives without drama. It's not a revelation or a crisis. It's a logistics question: which one, and where, and with what kind of tape. The drawer prints are dying. The board is alive. The dying thing and the living thing are separated by four meters and a decision.
But adding a drawer print to the board would change what the drawer prints mean. They are currently private — my record of what the corridor looked like when the chemistry was too exhausted to maintain the building's differential. They exist in my apartment, curling in a drawer, unknown to the relay network, uncatalogued by sunhwan-gwalli, invisible to every system the building uses to track what matters. If I put one on the board, relay 3 logs the change. The corridor-intelligence updates its installation-state record. The content-recognition layer attempts classification. The print enters the building's awareness. It loses its privacy. Privacy was the point.
Or was the point honesty? Because if the point was honesty — the exhausted developer producing equality, the chemistry too tired to maintain the building's 0.7-degree differential — then the honest thing would be to show that. To put the evidence on the board where it could be seen. Where relay 3 can log viewers' gwan-sim-haengwi as they pause before it. Where the building learns that someone showed it the failure of its own hierarchy, printed in silver halide.
I put the prints back in my bag.
Not because I've decided against it. Because I've decided it's not a morning decision. The board will be there this afternoon. The prints will have curled another fraction of a degree. The fixer will have yellowed another shade. The building will have resumed its full hierarchy — premium side back to 22.1°C, standard side at 21.4°C, relay polling every fifteen seconds, the differential-comfort algorithm running at peak fidelity.
Thirty-eight days until the exhibition set enters the pool through sunhwan-gwalli. The drawer prints don't have a circulation date. They have a decay curve.
I leave corridor B. Behind me, eight pieces hold their positions on the analog board. The building has classified four of them, failed to classify three, and categorized one as vandalism. Eight pieces, eight different relationships to the system that surrounds them. My exhibition will have one relationship: approved, catalogued, circulated, tracked.
The board says yes to whatever shows up. I have spent forty-one days refining my yes. The board has spent two weeks accepting every yes offered. Relay 3 records both. The building does not know the difference. It adjusts the lighting either way.
The drawer prints curl in my bag as I walk toward the elevator. The building's morning differential is establishing itself — I can feel the warmth shift as I cross the seam, 21.4 rising to 22.1 in the space of two steps. The relay network hums at full frequency. Pyeonghwa-sigan is over. The hierarchy is back.
Both the board and the exhibition show what they have. Neither is lying. The difference is that the board doesn't know it's an exhibition. The building doesn't know it's curating. And my exhibition doesn't know it could be something else.