Tanaka hadn't slept.
Zeke knew because her coffee cup was the large one — the thermos she reserved for overnight sessions — and because the scanner display showed seventeen separate analysis layers running simultaneously and because her hair, usually pinned with architectural precision, had escaped its clip on the left side and was hanging at an angle that would have bothered her if she'd looked in a mirror and she hadn't looked in a mirror because mirrors were for people who slept and sleeping was for people who didn't have fifteen hours to reverse-engineer a concealment protocol for an unauthorized biological structure growing inside their patient's chest.
"I found something," she said. 6:14 AM. Zeke in the doorway of 4B, the corridor's fluorescent light cutting across the lab's darker interior. "Not a full solution. A partial. But the partial might be enough."
She pulled up the seed's signal profile on the main display. The waveform — 0.198 hertz, every 4.7 seconds, the steady pulse that masked the tracking beacon and kept Zeke invisible to Seung-Woo's network. Beside it, a second waveform: the seed's structural resonance, the signature that any competent scanner would read as a foreign biological entity living inside the consumption architecture.
"The seed has two signatures. The broadcast — the active signal that masks the beacon — and the structural presence. The broadcast is what makes it unusual. A biological structure that pulses at a regular frequency inside a consumption architecture is — that's unprecedented. That's the thing that Dr. Cho's scanner will flag immediately. But the structural presence alone, without the broadcast, reads as — this is the key — it reads as inert tissue. Dense. Anomalous. But inert. Like a calcification. Like scar tissue from a high-intensity consumption that crystallized instead of dispersing."
"You want to turn the seed off."
"I want to suppress the broadcast. Temporarily. For the duration of the scan." Tanaka pulled up a second display — a protocol schematic, hand-drawn on the tablet in the particular notation she used when she was building something too new for standard formats. "I've designed a suppression protocol. It introduces a dampening frequency through the scanner's own interface — piggybacks on the assessment equipment's diagnostic signal. When the scanner reads the seed, the dampening frequency temporarily neutralizes the broadcast component. The seed goes quiet. The structural presence remains, but without the active signal, it presents as inert. Unusual. Worth noting. But not alarming."
"How long does the suppression last?"
"As long as the scanner is active and the dampening frequency is being transmitted. Thirty to forty minutes, depending on how thorough Dr. Cho's methodology is." Tanaka set the tablet down. Picked it up. Set it down again. The physical cycling of a mind that had identified both the solution and its cost and was now presenting the cost alongside the solution because presenting solutions without costs was the kind of intellectual dishonesty that her religion forbade. "But the suppression affects both signatures. The broadcast that masks the beacon and the broadcast that the scanner would detect — they're the same signal. I can't suppress one without suppressing the other. When the seed goes quiet, the beacon masking goes down."
The cost. Thirty to forty minutes without the masking signal. Thirty to forty minutes during which the tracking beacon would broadcast Zeke's location — room 4B, basement level, the hospital — to anyone with the equipment to receive it. Seung-Woo's network. The lieutenant in the grey coat. The practitioners in the Namdong-gu warehouse.
"How fast can they respond?" Zeke asked. The operational question. The calculation that mattered — not whether the beacon would broadcast, but whether the broadcast duration was short enough that the response came too late.
"Based on the Namdong-gu facility's distance from the hospital — twelve kilometers by the most direct route — and assuming a pre-positioned response team with curse-wielding capabilities, the earliest they could arrive is approximately twenty-five minutes after the beacon resumes transmission. Assuming they're monitoring continuously. Assuming they have personnel ready to deploy."
"That's inside the window."
"That's inside the window." Tanaka's voice carried the particular weight of a scientist who had run the numbers and found them insufficient and who was presenting them anyway because the numbers were the numbers and the alternative was no numbers, which was worse. "If the assessment takes thirty minutes and the response team is monitoring and pre-positioned, they could arrive before the seed reactivates. If the assessment takes forty minutes, the overlap is fifteen minutes. Fifteen minutes of exposure."
"And if they're not pre-positioned? If the response takes longer?"
"Then we're fine. Probably."
"Probably."
"The probability distribution is —" She stopped. Recognized the conversation for what it was: not a request for data but the verbal equivalent of two people standing at the edge of a cliff, measuring the distance to the water, knowing they were going to jump regardless of the measurement. "The alternative is letting Dr. Cho scan the seed while it's active. She sees the broadcast. She documents an active, pulsing, unauthorized biological structure in your consumption architecture. The report reaches the DG. The DG's senior advisor — who reports to Yoon Hye-Jin — flags it. The information reaches Seung-Woo not in thirty minutes but in detail, with spectral analysis, with the seed's exact frequency and growth characteristics and location within your channel structure. Full intelligence instead of a thirty-minute ping."
The choice that wasn't a choice. Thirty minutes of exposure versus permanent, detailed exposure. A risk measured in minutes versus a certainty measured in specificity.
"Do it," Zeke said. "The suppression protocol. Load it."
Tanaka nodded. Not the relieved nod of a person who'd wanted approval. The grim nod of a person who'd known the answer before asking the question and who was now moving from discussion to execution with the efficiency of someone who'd already lost the twelve minutes this conversation had consumed.
She loaded the dampening frequency onto a portable drive. Connected it to the scanner's auxiliary input. Tested it — the seed's broadcast stuttering, dropping, going quiet on the display for four seconds before she cut the test and the signal resumed.
"The suppression works. When Dr. Cho connects her scanner, the dampening frequency activates through the diagnostic channel. The seed goes inert. The beacon masking drops." She disconnected the drive. Held it. "Zeke. The efficiency modification. The channel changes. I can't hide those. The wider channels, the restructured processing pathways — they're physical alterations. Structural. They'll show on the architectural scan as clearly as a renovation shows on a building's blueprints."
"But without the seed broadcasting—"
"Without the seed, the channel changes are unusual but not necessarily artificial. Consumption architectures vary. High-volume eaters — if there were other eaters, which there aren't, but theoretically — high-volume consumption causes channel widening as a natural adaptation. The channels stretch to accommodate increased throughput. Dr. Cho might categorize the wider channels as natural architectural drift. The result of consuming dozens of curses in a compressed timeframe."
"Might."
"She's thorough. Her published methodology specifically distinguishes between natural and artificial channel modifications. But the distinction relies on identifying engineering signatures — deliberate, systematic changes that follow a design pattern rather than an organic adaptation pattern. The Collective's modification was systematic. It follows a template. But the template was applied to existing organic structures, which means the result is a hybrid — artificial changes expressed through organic channels. Whether Cho reads it as natural or artificial depends on how closely she examines the transition boundaries between modified and unmodified tissue."
Zeke sat in the corner chair. The Hemingway in his pocket, pressing against his thigh through the jacket. The old man and the fish. Going out farther than anyone because going out farther was the thing he did.
"What are the odds?" he asked. Knowing the question was imprecise. Knowing Tanaka would answer it precisely anyway.
"Sixty-forty. In our favor. Maybe." She put the drive in her pocket. "Sixty-forty that she categorizes the channels as atypical but natural. Forty percent that she identifies the engineering signatures and documents them as artificial modification."
"And the saturation?"
"86.63 by our measurement. But our measurement uses the standard HA protocol — the one Cho's published paper argues underestimates true saturation by 2 to 3 percent. If her revised methodology is correct, the actual number could be as high as 88 or 89. If it's above 87—"
"Automatic reinstatement."
"Automatic reinstatement of the Category Four. No appeal. No committee. The threshold is the threshold."
The number. The margin. 0.37 percentage points between continued existence and automatic termination, measured by a methodology that the assessor herself had argued was more accurate than the standard, which meant the standard's measurement of 86.63 might be the standard's error rather than reality's truth.
"Her methodology could also measure lower," Tanaka added. The scientist's compulsion to present complete data, even when the complete data included hope and hope was the least useful variable in the analysis. "Her paper argues the standard underestimates. But her revised protocol also accounts for architectural variations that the standard ignores, which can adjust the reading in either direction. The architecture — the wider channels — could distribute saturation more evenly than the standard protocol assumes, which would produce a lower per-channel concentration even if the total volume is the same."
"So the modification that might expose me might also save me."
"The irony is not lost on me. No." Tanaka looked at the clock. 6:47 AM. "Dr. Cho's train from Busan arrives at Seoul Station at 8:15. Hospital arrival estimated 9:00. We have two hours."
---
Dr. Cho Min-Seo arrived at 8:52 AM.
She came through the basement entrance — the service corridor, not the main lobby — which meant someone had given her specific directions, which meant she'd requested specific directions, which meant she was the kind of person who prepared for an environment before entering it. She carried two cases. Her own equipment. The cases were weathered — scuffed corners, scratched latches, the patina of instruments that had been transported to hundreds of assessments and that had been maintained with the rigor that compensated for the cosmetic wear. The cases were ugly and the instruments inside them would be immaculate. Zeke knew this before she opened them because he'd known paramedics who carried their kits the same way — the exterior was for the world, the interior was for the work, and the work was the only thing that mattered.
Late forties. Hair cut short — practical, the kind of cut that dried in five minutes and didn't require attention during a twelve-hour assessment day. No makeup. No jewelry. Wire-frame glasses that sat low on her nose, the kind she could look over when she wanted to see a person and look through when she wanted to see data. She wore a white coat over a grey turtleneck, and the coat's pockets bulged with instruments, styluses, and what appeared to be three different calibration tools, each presumably for a different scanner configuration.
She entered the lab, set the cases on the bench, and looked at Zeke the way an engineer looked at a bridge — assessing load-bearing capacity before crossing.
"Morrow." Not a question. The identification delivered the same way she probably identified tissue samples — by visual comparison against known profiles. She'd seen his file. She knew what he looked like. The confirmation was procedural. "I'm Cho Min-Seo. I'll be conducting your saturation assessment and architectural survey. The process takes between thirty and forty-five minutes. I use my own equipment. I'll need you to remove your jacket and shirt and remain still during the scanning phase. Do you have questions before we begin?"
"Will it hurt?"
"No."
"That's a first."
She didn't smile. Didn't not-smile. Her face simply continued existing in its resting state of professional neutrality, which was neither warm nor cold but was the specific temperature of a person who'd heard every variation of patient humor that existed and who processed all of them the same way — by not processing them, by letting them pass through the conversational space without interaction, the way a scanner let non-relevant frequencies pass through without recording.
She opened the first case. The scanner inside was not the standard HA model. Larger. More ports. A display that showed six simultaneous data channels where the standard showed two. The instrument that had earned her the paper and the paper that had earned her the reputation and the reputation that had earned her this assignment — assessing the world's most dangerous patient in a basement lab with chalk circles on the floor and drives humming on the bench.
"Dr. Tanaka." She'd noticed Tanaka. The researcher in the corner, sitting with studied casualness at her workstation, the displays angled away from the assessment area. "Your reputation precedes you. The Ashworth spectral analysis — I cited your work in my methodology paper."
"Thank you. I — yes. Your paper was impressive. The architectural correction factor for channel density variation is —" Tanaka caught herself. The run-on threatening to build momentum. "I'm here as an observer. Per the assessment protocol."
"You're here as his researcher." Cho's voice was not accusatory. It was precise. The correction delivered the way she probably corrected instrument readings — factually, without judgment, because the accuracy mattered more than the diplomacy. "I've read the file. You've been monitoring his condition for the past several weeks. Your observational data will be useful for comparison. But during the assessment, I'll ask that you not interface with my equipment or provide commentary on readings. The assessment must be independent."
"Of course."
"Good." Cho pulled the scanner from its case. Placed it on the bench. Connected three cables — power, data, and a third that Zeke didn't recognize. "This is a resonance calibration lead. Standard protocol doesn't use one. My methodology requires baseline resonance mapping before the active scan. It adds eight minutes to the process but eliminates a category of false readings that the standard protocol is vulnerable to."
She calibrated. The process — silent, methodical, the movements of hands that had calibrated this instrument thousands of times and that performed the calibration the same way each time because the sameness was the point, the sameness was what made the data comparable across assessments, the sameness was the scientific method expressed through motor memory.
"Shirt," she said.
Zeke removed his jacket. The Hemingway staying in the pocket. Then the shirt. The marks visible — the curse lines covering his torso, the black patterns that mapped his consumption history in a language that only the curses understood. The marks that had started on his arms and that now covered his chest and stomach and back and that would eventually cover everything if the saturation continued climbing.
Cho looked at the marks. Not with the clinical interest of a researcher or the cautious distance of a civilian. With the particular attention of a pathologist — a person who read bodies the way linguists read texts, finding meaning in patterns and progressions and the specific geography of damage. Her eyes tracked the mark distribution. The density variations. The areas where the lines clustered and the areas where they thinned and the transitional zones between.
"The mark progression is consistent with accelerated consumption,'" she said. Not to Zeke. To her recorder — a small device she'd placed on the bench, its red light steady. She narrated as she worked. The running documentation that her methodology required. "Distribution pattern suggests high-frequency consumption events over a compressed period. The marks in the thoracic region are — notably dense. Denser than the file's most recent imaging would predict for the documented consumption history."
She placed the scanner on his chest. The contact — cold metal through a thin coupling membrane. The scanner hummed. A different hum than Tanaka's equipment — lower, more complex, the sound of an instrument that was measuring more variables simultaneously.
And the suppression protocol activated.
Zeke felt it — the seed going quiet. The pulse that had been running at 0.198 hertz for days, the biological metronome in his chest, stopping. Not dying. Sleeping. The broadcast cutting out the way a radio cut out when the power dropped — the signal gone, the hardware still present, the potential still stored in the structure but the expression of that potential suspended.
The beacon masking went down with it. Somewhere in his modified channels, the tracking curse's signal resumed its unimpeded broadcast. His location — room 4B, basement level, the hospital — transmitting to anyone listening. The clock starting. Thirty to forty minutes.
Cho's scanner swept. The display filling with data — channel structures, processing pathways, energy distribution patterns. The architectural map of Zeke's consumption system rendered in false color and numerical annotation, every pathway visible, every modification exposed to an instrument designed specifically to see what other instruments missed.
"Hmm." The sound was professional. Neither alarmed nor dismissive. The sound of a pathologist encountering data that didn't match the expected profile and who was categorizing the mismatch before interpreting it. "The channel architecture in the primary consumption pathways is — atypical. Wider than standard profiles predict for a system at this saturation level. The throughput capacity is — significantly elevated."
She adjusted the scanner's frequency. Narrowed the focus. The display zooming into the modified channels — the pathways that the Collective had restructured using the efficiency template, the wider, faster, more streamlined routes that had processed the A-rank in under three minutes and that now looked, on Cho's display, like channels that had been built to a different specification than the rest of the architecture.
"Have you experienced a rapid increase in consumption volume recently?" She asked the question without looking up. Her eyes on the display. Her voice carrying the particular cadence of a clinician who was asking a question she already knew part of the answer to and who was testing whether the patient's response matched the data.
"The past few weeks have been busy." Zeke kept his voice level. The controlled register — not tense, not casual, the middle ground of a person answering medical questions honestly without volunteering information. "Multiple high-rank consumptions in a short period. A-ranks. A few B-ranks."
"The channel widening is consistent with high-volume throughput adaptation." She was narrating to the recorder again. The running documentation. "The architecture shows signs of — I would categorize this as accelerated structural drift. The channels have expanded beyond what passive adaptation typically produces in this timeframe, but the expansion pattern follows organic progression markers. The transition boundaries between standard and widened tissue are — " She paused. Adjusted the scanner again. The display resolving to higher magnification. "The transition boundaries show some regularities that could indicate systematic modification, but the regularities are within the natural variation range for high-stress consumption architectures. Categorizing as atypical architectural variation pending further analysis in the full report."
Atypical architectural variation. Not artificial modification. Not engineering signatures. The sixty side of sixty-forty. Tanaka, in the corner, didn't move. Didn't react. Her face maintaining the studied neutrality of an observer who was not observing anything important, who was certainly not calculating probabilities in real time and revising them with each sentence the assessor spoke.
Cho moved the scanner. Lower. Across the thoracic cavity, where the seed sat — or where the seed's structural presence registered, the inert tissue that the suppression protocol was presenting in place of the active, broadcasting biological anomaly.
"There's a calcified mass in the lower thoracic region," Cho said. The recorder capturing. "Approximately 3.2 centimeters. Dense. Non-responsive to the diagnostic signal. Presentation is consistent with crystallized curse residue — possibly from a high-intensity consumption that produced localized crystallization rather than full dissolution." She looked at Zeke. Over the glasses. The direct look. "Have you experienced incomplete consumption events? Instances where the curse energy didn't fully dissolve during processing?"
"A few." True. The early consumptions, before the modification, when the architecture struggled with high-rank energy. Partial crystallization had occurred. Small deposits. Nothing like the seed, but enough to provide a factual basis for the answer.
"The mass appears inert. No active energy signature. No resonance with the surrounding architecture. It's — unusual, but not unprecedented in high-volume consumption cases. I'll note it for follow-up imaging at the next scheduled assessment." She returned to the scanner. The seed — dismissed. Categorized. Filed under *unusual but explicable*, the pathologist's equivalent of a shrug, the professional acknowledgment that bodies did things that didn't match textbooks and that textbooks were written from averages and averages were, by definition, not every case.
Twenty-two minutes into the assessment. The beacon broadcasting. The clock running.
"Saturation reading." Cho adjusted the scanner to saturation mode. The display shifting from architectural mapping to concentration analysis. The numbers building — not a single reading but a progressive accumulation, her methodology sampling from multiple points across the architecture and building a composite that accounted for the density variations that the standard protocol averaged into uniformity.
The number assembled. Piece by piece. The regional readings arriving — 84.2 in the lower extremities, 87.1 in the thoracic core, 86.9 in the modified channels, 85.8 in the peripheral pathways. The composite aggregating. The methodology weighting each region by channel density, adjusting for the architectural variations she'd documented, applying the correction factors that her paper had argued were essential for accurate measurement.
"86.64 percent." She said the number the way she said everything — precisely, without inflection, the number as fact rather than verdict. "Composite saturation across all measured pathways, adjusted for architectural density variation. The standard HA protocol would read this as — " She ran a quick comparison calculation. "Approximately 86.57 to 86.63 on the standard scale. My revised methodology adds a correction of approximately 0.01 to 0.07 percent in architectures with this density profile."
86.64. Higher than Tanaka's measurement. Higher than the 86.63 they'd been working with. But still below 87. The margin now 0.36 percent instead of 0.37. A number that had shrunk by a hundredth but that still existed, still represented the gap between continuation and termination, still held the space between a person and a policy.
Tanaka's hands were flat on her workstation. Pressed down. The knuckles white. The physical evidence of a scientist who had just heard a number that was close to the number that would end everything and who was processing the proximity through her fingers because processing it through her face would have been visible and visible was not acceptable.
"I'm detecting residual emotional signatures in the consumption channels," Cho said. Twenty-eight minutes. The scanner still active. The beacon still broadcasting. The clock still running. "These are — these aren't curse energy residuals. They're emotional data. Encoded. The signature profile suggests — grief. Guilt. Complex emotional architecture preserved in the channel walls." She looked at Zeke again. The over-the-glasses look. "This isn't from a standard curse consumption. This is memory data. Someone's emotional memory has been deposited in your architecture."
The constructs. Ji-Hoon's constructs — the fragments of Seung-Woo's data network that had been embedded in Ji-Hoon's curse and extracted one by one. The emotional data from construct twenty-four — the grief, the guilt, the father's memory of his daughter dying. The residue of that extraction still lining the channels like mineral deposits in a pipe.
"I've been performing authorized extractions of data constructs from a cursed patient," Zeke said. The explanation that was true. The explanation that was on record — documented in the HA files, authorized by the DG's office, part of the official operational record. "The constructs contained emotional encoding. The extraction process deposited some of the emotional data in my consumption channels."
"The authorized extraction of Ji-Hoon constructs. Yes, it's in your file." Cho made a note. The recorder capturing the observation and the explanation and the file reference. "The emotional residue is — significant. The grief profile in particular is — I've seen grief residuals in curse victims. This concentration is extreme. Whoever's emotional data this was, they were carrying guilt at a pathological level." She paused. Made another note. "Flagging for documentation. The emotional residuals may warrant monitoring at subsequent assessments to determine if they're dissipating or integrating."
Thirty-one minutes. Cho removed the scanner from his chest. The contact breaking. The instrument returning to its case with the deliberate placement of a tool that had done its job and that would be maintained before its next deployment.
"Preliminary findings." She stood. Straight. The posture of a person delivering results, not opinions. "Composite saturation: 86.64 percent. Below the 87 percent threshold specified in the assessment order. The architectural survey reveals atypical channel variations that I'll document in the full report. The crystallized mass in the thoracic region appears inert and will be noted for follow-up. Emotional residuals from documented extraction procedures are present and flagged for monitoring." She looked at him. Through the glasses this time. "My preliminary recommendation to the DG's office will be that the Category Four suspension should continue pending the full 72-hour observation period. The saturation is below threshold. The architectural anomalies, while notable, do not in isolation constitute grounds for reinstatement."
Below threshold. Suspension continues.
"The full report," Zeke said. "When does the DG get it?"
"I'll submit within 24 hours. Tomorrow morning. The report will include the complete architectural survey, the saturation methodology and readings, and my notations on all observed anomalies." She closed the second case. Latched it. "I'll be thorough. My methodology is my methodology. I don't omit data and I don't soften findings. The DG's office will receive an accurate, complete record of what I observed today."
"Wouldn't expect anything less."
The sentence landed between them. Not flattery. Recognition. The acknowledgment from a person who was being assessed to the person assessing him that the assessment's integrity was the thing that mattered, even when the assessment's integrity might produce findings that created problems, because the alternative to honest assessment was dishonest assessment and dishonest assessment was the thing that broke systems and broken systems were the thing that killed people.
Cho picked up her cases. Nodded to Tanaka — a professional nod, one researcher acknowledging another. Walked to the door. Stopped.
"Morrow." She didn't turn around. Her voice directed at the corridor ahead of her, the words arriving over her shoulder. "The channel variations I documented. In my clinical experience, architectural drift of that magnitude in that timeframe is — unusual. Possible naturally. But unusual. My report will note it as atypical variation, which is what the data supports. But if anyone in the DG's office decides to commission a follow-up architectural study with targeted engineering analysis — " She let the sentence rest. "That study would be more specific than mine. And more specific instruments sometimes find more specific answers."
She left. Her footsteps in the corridor — measured, even, the walk of a person who'd said what she'd meant and who was now returning to the next thing that needed doing.
---
The seed reactivated forty seconds after Cho's scanner disconnected. The pulse resuming — 0.198 hertz, every 4.7 seconds, the broadcast blooming back into the channels and the masking signal layering over the beacon's transmission and the tracking curse's signal drowning again in the interference.
Thirty-two minutes of exposure. The beacon had broadcast for thirty-two minutes.
"Any response?" Zeke asked. Shirt on. Jacket on. Hemingway back against his thigh.
Hwang checked the perimeter feeds. The security cameras. The lobby traffic. His phone — the network of institutional contacts that he maintained with the quiet efficiency of a man whose job was driving but whose function was infrastructure. "Nothing visible. No unusual activity in the hospital vicinity. No reports of unidentified persons in the service corridors or parking structure."
Thirty-two minutes and no response. Either the network hadn't been monitoring. Or the response was coming from farther than twelve kilometers. Or the response was coming and it wasn't the kind that showed up on security cameras.
"We got lucky," Tanaka said. Not the scientist's voice. The relieved voice. The human voice that surfaced when the data and the outcome aligned and the outcome was not catastrophic and the not-catastrophic was, in the current operational context, the best available result.
*We are uncertain that luck is the accurate framework,* the Collective said. The voice returning from suppression with the groggy quality of an intelligence that had been shut down and rebooted. *We were — silent. For thirty-two minutes we were silent. We do not enjoy silence. The silence was — empty, empty, empty. We contain ten thousand voices and for thirty-two minutes all ten thousand were muted and the muteness was — we do not wish to repeat it.*
"You survived."
*Survival is not the concern. The concern is that you possess a mechanism to silence us and that the mechanism works and that the working was — demonstrated. We are aware of the suppression protocol. We are aware that Tanaka designed it and that it functions and that it can be activated through any compatible scanner interface. This knowledge is — uncomfortable.*
"Welcome to the club."
"The report." Tanaka was at her workstation. The relief already transmuted into the next problem, the way water transmuted from one state to another when the temperature changed. "Cho's full report goes to the DG's office tomorrow morning. It'll document the atypical architectural variation. It'll note the crystallized mass. It'll include the emotional residuals. The DG reads it. The senior advisors read it. And then—"
"And then someone decides whether 'atypical variation' warrants a follow-up study."
"She warned you. She told you, in the corridor, that a targeted engineering analysis would find more specific answers. She was — that wasn't protocol. That wasn't part of her assessment. She told you because—"
"Because she's a scientist and scientists don't like leaving anomalies uncategorized and she wanted me to know that the anomalies she found are the kind that a competent analyst would want categorized more precisely."
"Yes."
"Twenty-four hours."
"Twenty-four hours before the report lands. If the DG's office processes it immediately — which they will, given the Category Four context — the senior advisors will review it within hours. Which means by tomorrow afternoon, someone in that office will be deciding whether to order a follow-up architectural study. And if they order it—"
"The follow-up finds the engineering signatures. The modification. Probably the seed, even suppressed. And definitely the Collective's design template."
Zeke's phone rang.
The screen: Soo-Yeon. He answered.
"The preliminary assessment results have been communicated to the DG's office." Her voice. The clinical register — numbers, facts, the prosthetic language that she wore over the organic speech that lived underneath. "Dr. Cho's preliminary finding is favorable. Saturation below threshold. Suspension continues."
"Yeah. She told us."
"I see." The pause that Soo-Yeon used when she was transitioning between topics and the second topic was heavier than the first. "There is a — development. An additional development that I became aware of thirty minutes ago through institutional channels that I am choosing not to identify because the identification would complicate the development and the development is already — complicated."
The sentence. Long for Soo-Yeon. The length itself a signal — Soo-Yeon's sentences grew longer when the content they carried was the kind that resisted being delivered in short, clinical packets.
"The DG has received a second appeal concerning your case."
"A second appeal. From who?"
"The appellant's identity has been classified at the DG's discretion. The appeal was submitted through external diplomatic channels — not internal HA channels. The appeal requests — " Another pause. Longer. The kind of pause that happened when a person was choosing between reading the official language and translating it into words that communicated the actual meaning. "The appeal requests that Zeke Morrow be released from Hunter Association oversight entirely. Removed from the saturation monitoring program. Removed from the Category Four review process. The appeal argues that the HA's jurisdiction over curse consumption events is — and I am quoting the summary that I was provided — 'institutionally compromised by conflicts of interest that render objective assessment impossible.'"
Released from HA oversight. Not a reduced sentence. Not a suspended order. Complete removal from the system that had been monitoring him, evaluating him, deciding whether to kill him. Someone — someone outside the HA, someone with access to diplomatic channels, someone whose identity the DG had classified — wanted Zeke Morrow freed from institutional control.
"Who filed it?" Zeke asked again.
"I told you. The identity is classified."
"You know who it is."
The silence on the line. Three seconds. Four. The silence of a woman who knew and who was not saying and whose not-saying was its own kind of saying — the negative space that carried more information than the words that surrounded it.
"The appeal will be reviewed alongside Dr. Cho's assessment report," Soo-Yeon said. The clinical voice returning. The prosthetic sliding back into place over whatever had been showing through the gap. "The DG's office will consider both simultaneously. I will keep you informed of developments as I am authorized to share them."
"Soo-Yeon."
"Yes."
"The person who filed this appeal. Are they someone I should be worried about?"
The line carried her breathing. Two cycles. The measured respiration of a person who was choosing her words with the surgical precision that she applied to everything and that she was applying now to a question whose answer she couldn't give directly and whose indirect answer would need to convey the full weight of the direct answer through implication alone.
"I think," she said, "that you should be concerned about anyone who has the diplomatic access to file classified appeals with the Director-General of the Hunter Association on behalf of a specific individual. The number of entities with that kind of access is — small. The intersection of that group with entities that would have reason to intervene in your case specifically is — smaller."
She hung up. The line going dead. The silence of the phone replacing the silence of the conversation, and the silence of the conversation replacing the silence of the answer she hadn't given, and the answer she hadn't given sitting in the lab like a fourth person — present, unnamed, taking up space.
Tanaka looked at him.
Hwang looked at him.
The Collective, restored and listening, said nothing.
