A busy weekend (revised)
Cory peeled off her work gloves. Nitrile, single-use, heavy-duty. Tossed them in the recycling chute. Missed. Picked them up. Tossed again. Her hands were trembling slightly, the fine-motor tax of thirty-six hours on her feet.
The crew shuttle had left twenty minutes ago. She should have left with them. The rest of the eight-person Altiva crew was already halfway back to Monterey, packed into a crammed vehicle that smelled like sweat and aerogel adhesive, probably falling asleep. She had waved them off. She always stayed last.
Madison Heights. One of the very first second-gen installations: a five-story office complex commissioned in 2050, the year the Tesseral Materials certification board approved EuGold® anti-grav panels for commercial use. Twenty-eight years afloat. The residents' association kept it immaculate: fresh aerogel insulation, upgraded fiber optics, the kind of maintenance work that clients actually enjoyed paying for. Cory appreciated that. She hated being asked to cut corners.
She stepped onto the exterior ramp and into the cold. The construction lights had already powered down, and the ramp's red markers blinked in slow rhythm, barely enough to see by. The mountains to the east were just visible against the last light of dusk, dark shapes with snow-lit ridges, impossibly distant, impossibly below. She was a little over a thousand meters up. The parking ramp hung off the building's north face on steel cables that creaked when the temperature dropped, which it was doing now.
She put her hand flat on the polished stone of the exterior wall and held it there.
This was her ritual. Every job, last thing: the final check, conducted with no instrument but her palm. Eight years of structural work had trained her, just like decades of tuning pianos train a professional ear. Telemetry gave her numbers. Her hand gave her the thing the numbers were about.
The building hummed. Normal. The deep, almost subliminal vibration of four thousand EuGold panels holding twelve thousand tons motionless in mid-air. She knew this hum the way she knew her own pulse.
But under the hum, something else. A faint lateral unsteadiness, like a gyroscope that has begun, just barely, to precess. Two seconds. Then gone.
She frowned. Stood still. Eyes closed, fingers spread against the cold stone, waiting for it to come back. Buildings settle after renovations, that's normal, expected, but always vertical. Never lateral. Lateral made no sense.
She waited for a long time. The wind picked up slightly, and the parking ramp swayed on its cables. Nothing from the wall.
Then she stopped herself, because she'd been awake for thirty-six hours and her mind was blurring at the edges. She knew what that meant. Not that she was wrong. But that she couldn't trust the next thought to be as reliable as the last one.
She tapped her wrist and called Rua.
Rua: "Vasquez. You're supposed to be off-site."
Cory, correctly, read this as surprise, not accusation.
Cory: "Job's done. But there's something wrong with the building. I felt an anomaly."
A short pause. Somewhere below, a driverless cargo truck quietly floated by, its lights briefly illuminating the ramp.
Rua: "Wrong how?"
Cory: "Lateral instability. Subtle. I can feel it but I can't quantify it. Might show up in the thruster logs, they could be compensating more than usual."
Rua: "If I send someone else out there tomorrow, will they feel it?"
Cory: "Probably not."
Rua: "But you did."
Cory: "Yeah."
Rua: "Go home. Sleep for two days. I mean that, two actual days, not five hours and a protein bar. I'll pull the maintenance logs."
Cory: "Rua."
Rua: "I heard you. I'll deal with it. I don't want to hear from you before Thursday."
Cory hung up. She stood on the ramp for another moment, looking out at the dark. Three other buildings floated in the middle distance, window-light sharp against the sky. She'd helped build one of them. The cables creaked. The red markers blinked. Twelve thousand tons of architecture, silent and motionless above a world that had mostly stopped wondering about how strange that was.
Her glider unlocked automatically as she approached. She dropped into the single seat, pushed start, and the vehicle pulled away into the dark. Madison Heights shrank behind her, just another lit rectangle against the mountains, solid as anything.
She scrolled through music tracks. Couldn't find the right one. Gave up and flew home in silence.
---
Cory slept for fourteen hours, woke up, ate standing at the kitchen counter, reheated soup, the spoon clinking too loud in the quiet apartment, and slept for ten more. The altitude work took its toll.
Meanwhile, Rua got busy.
The maintenance logs for Madison Heights told a quiet story, if it was a story at all. The thrusters' energy consumption had been trending upward for two years: too slight to trigger an automated flag, too gradual to show in any single month's numbers, visible only as a faint shape when you plotted it month by month and wanted to see it. Rua plotted it. Then she pulled logs from the other aging buildings Altiva had serviced recently: Coronado Terrace in San Diego, Bellevue Seven above the strait in Seattle, and Park Tower Kamakura, which they'd consulted on for a Japanese firm. All second-gen. All between twenty-five and thirty years old.
She plotted those too. The curves were noisy, genuinely noisy, full of seasonal spikes, sensor recalibrations, maintenance windows that broke the trendline. Individually, each building's data could mean anything or nothing. Plotted together, all four showed what might be the same shape underneath: a slow, monotonic climb in thruster compensation, steeper in the older buildings.
A statistician would have run tests on that data and concluded that the upward trend was indistinguishable from random noise.
Rua was not a statistician. She was a site manager who had spent twenty years paying attention to what her crew told her. Her best crew member was telling her something was wrong, and the numbers were not telling her it wasn't.
She called Tesseral Materials from her office in the Altiva building in Monterey, a building that was itself floating sixty meters above the old cannery district on second-gen EuGold panels. The call connected to an AI intake agent: polite, calibrated, working through a diagnostic script. No, the panels did not get damaged during shipping. Was this a warranty issue? Let's say yes. The agent offered to schedule a standard panel diagnostic through her regional service center. Rua declined and asked for a human specialist. The agent offered again, reworded, patient. Rua declined again. Eleven minutes of increasingly pointed conversation with a system designed to resolve everything at the lowest possible level before it finally conceded that this was beyond its resolution authority and connected her to a man who introduced himself as Walter Lam.
Walter: "Thruster usage increases for a dozen reasons," Walter said. He had the clipped efficiency of someone who fielded these calls regularly. "Environmental exposure, sensor degradation, seasonal load changes. What you're describing is consistent with normal fleet variation."
Rua: "I've got four buildings in three climate zones all trending the same direction."
Walter: "With four data points and this noise level, you'd expect apparent trends by chance. Small sample, high variance. If you ran the same analysis on four randomly selected buildings of any age, you'd find a similar shape. It's not statistically meaningful."
Rua: "The older ones are climbing faster."
Walter: "Or the noisier ones happen to be older. Age correlates with legacy sensor packages, older actuator models, different maintenance protocols. You're pattern-matching across confounds." A short breath, not quite a sigh, the sound of someone preparing to close a conversation cleanly. "Ms. Rua, if you do observe instrumented panel-level response degradation, you're welcome to submit it through your regional service center and we'll review it. Otherwise, I'd recommend a standard panel diagnostic. I can have the intake system send you the scheduling link."
Rua: "One of my crew felt a lateral instability on Madison Heights. Hands on the wall. She says—"
Walter: "I appreciate you calling, and I don't doubt your crew's experience. But Tesseral monitors six hundred installations worldwide, continuously, with dedicated sensor arrays. If there were a systemic material trend, we'd have seen it. I'd encourage you to schedule that diagnostic. Is there anything else I can help with today?"
There wasn't. She could hear the file closing.
Rua: "No. Thank you for your time."
She sat back. Stared at the ceiling of her floating office.
He wasn't wrong. That was the problem. The evidence wasn't strong enough to prove what it hinted at. But the telemetry data was rough, imprecise. Cory's intuition was not. Rua had built her career by relying on her workers' intuitions. And she'd just had a door shut in her face by someone who hadn't even been curious enough to look at the numbers.
She needed someone who could look at the material itself.
She opened a search window and pulled up Cory's publication record. She'd read it three years ago, after the conversation about overqualification. The site lunch where she'd told Cory she was the most overqualified person Rua had ever managed and that it was a damn shame. She'd looked up the papers afterward. Not because she understood ceramic quasi-crystallography, but because she believed in knowing what her people could do. All of it. Especially the parts they'd stopped using.
The dissertation sat at the top of the record: Vasquez, C. (2070). Phase Boundary Migration in Quasi-Crystalline Ceramics Under Sustained Thermal Cycling. Doctoral thesis, UC Santa Cruz.
One name appeared on two of the associated journal articles. Same name on the faculty page of NTU Singapore's materials science division: Priya Chandrasekaran. Fatigue testing. Long-duration cycling behavior in engineered ceramics. Active lab, active publications, currently leading a research group of eleven.
Rua tapped the contact link on Priya's faculty page at nine in the morning Monterey time. Midnight in Singapore. Priya answered on the third ring, sounding neither sleepy nor surprised. Rua laid it all out: the four buildings, the thruster trends, the noise, Walter's rejection. She said one of her crew had felt a lateral instability on a twenty-eight-year-old second-gen building. She said the crew member's name was Cory Vasquez.
Then she asked if Dr. Chandrasekaran would be willing to look at the data independently and tell her whether there was anything there.
A pause. Then a hum on the line that Rua couldn't interpret, but that sounded like the audible version of someone leaning forward.
Priya: "Yes. Send me everything."
Rua sent the files and went home. Past eleven. She'd done what she could do for today.
---
Priya called Cory the next afternoon. Cory was on her couch with reheated soup and a body still settling back toward equilibrium. Twenty-four hours of sleep had taken the edge off, but her joints ached and her hands still had the dry, cracked texture of three weeks at altitude. Her wrist pulsed. A name she hadn't seen in over two years. They kept up the way grad school friends do when life pulls them into different orbits — a paper forwarded with a one-line note, a birthday message, a photo of Priya's new lab with no caption needed. But they hadn't actually spoken since the call when Priya made associate professor, and even that had been brief, warm, and interrupted by someone knocking on Priya's office door.
Priya: "Hi. Cory."
The voice was exactly as she remembered, mid-thought, full velocity, arriving like a sentence already in progress. Two years of silence and the shorthand was still there, immediate, like a language the body doesn't forget.
Cory: "Priya?"
Priya: "Your boss called me last night."
Cory sat up. Soup sloshed.
Cory: "My boss?"
Priya: "Rua. Operations director, Altiva Structural. She found me through your papers, looked up your publication record and tracked me down. Called me immediately and explained everything without being suggestive, which I respect. She said you felt something wrong with a building and that she believes you."
Cory didn't speak. She was holding two things at once: the surprise of hearing Priya's voice, and the quieter thing underneath, the realization of what Rua had done. Rua had gone looking. Had read her papers. Found a name. Picked up the phone at midnight and asked a stranger for help, because Cory's hand on a wall had told her something was wrong and she didn't have the tools to prove it.
Cory had worked for Rua for six years. Rua had never once asked about her PhD, never referenced her papers, never treated her academic past as anything other than a fact she'd noted and filed away. But she had filed it. And now she'd pulled it out.
Priya: "Cory? Are you there?"
Cory: "I'm here. Tell me what you know."
Priya: "The thruster curves. Individually, yes, they're a mess, I understand why Tesseral dismissed them. But the composite shape, Cory. There's something under the noise, and I want to know if it's what I think it is, because if it is, we have a problem."
Cory: "What do you think it is?"
Priya: "Boundary migration. Your dissertation mentioned it as a theoretical possibility, but we never tested for it. If it's real, the panels are drifting under long-term thermal cycling and the self-correction is starting to fall behind. That's what the thrusters are compensating for."
Three seconds of silence. For Priya, a geological epoch.
Then the hum, low, sustained, the sound she made without realizing it when she was circling something important.
Priya: "I just realized I have samples. Offcuts from one of the first manufacturing runs. Tesseral originals, second gen. I got them for a different project three years ago. I can run accelerated thermal cycling, compare the grain boundary profiles against your predicted drift curves. If the boundaries are moving the way you described in your 2070 paper, I'll see it in the lattice. Directly."
Cory: "How long would that take?"
Priya: "A week. Maybe less. I'm canceling the ICMC trip, this is more important." A pause. The hum. "Cory. If you're right, this is happening in every second-gen building in the world."
Cory: "I know."
Priya: "And nobody's seen it because nobody's been looking on this timescale."
Cory: "Because the timescale hadn't elapsed yet. We certified the material based on the best data available, which wasn't enough, because it couldn't be."
Priya: "Nobody did anything wrong and the problem is still real. I hate this kind of problem." The hum again. "I love this kind of problem. Send me everything. Including your dissertation."
Cory: "You already have my dissertation."
Priya: "I want to read it again. I want to read it the way it should have been read in 2070."
Cory hung up and sat for a while, one hand resting on the other wrist where the display had already gone dark. Through the window she could see the underside of the Salinas Residential Platform, twenty-two years old, second-gen, budget-series, home to about nine thousand people. She'd looked up at it every morning for four years. The shadow it cast moved across the neighborhood in a slow daily arc, and she'd arranged her couch to catch the afternoon sun that came through when the platform drifted west. She knew its schedule like a neighbor's.
The familiar view had lost some of its comfort.
She called Rua.
Rua answered on the second ring.
Cory: "You read my papers."
Rua: "I read everybody's everything."
Cory: "Come on."
A short pause. When Rua spoke, her voice had its usual clarity, level, direct, but underneath it, briefly, a warmth she usually distributed in smaller doses.
Rua: "You wrote about phase boundary migration in quasi-crystalline ceramics under thermal stress. I didn't understand all of it, but I understood enough to know it was about what happens to the material over time. And now I've got four buildings telling me something is happening to the material over time. So I went looking for someone who could help. You put her name on your papers. I followed it."
Cory, quietly: "Thank you."
Rua: "Don't thank me. Get ready. Priya says she'll have results soon. You'll need to be sharp when she does."
Cory: "Rua. My rest days end tomorrow. I need to go back on rotation."
Rua: "No you don't."
Cory: "I can't just sit here and wait for Priya's results on unpaid leave. Rent's due in—"
Rua: "I put you on Altiva's technical consulting rate this morning. Backdated to the night you called me from Madison Heights."
Cory didn't say anything.
Rua: "Every aging building we've serviced is a liability exposure for this company if what you and Priya are looking at turns out to be real. I explained that to management and they saw the logic quickly. You're on this full time, Cory. Paid."
Cory: "That's generous, but—"
Rua: "It's not generous. It's resource allocation. You're the most qualified person I have for this problem and I'm not sending you to go re-insulate somebody's atrium while we wait to find out if the sky is falling. Are we clear?"
Cory: "We're clear."
Rua: "Good. I'll send you the paperwork. Now go eat something that isn't reheated soup."
---
Priya's results came back in five days. The grain boundaries in her archived samples showed exactly the micro-rotation Cory's dissertation had predicted, subtle, cumulative, consistent with the thermal history of panels in service for decades. The confirmation was clean, unambiguous, and frightening.
Cory called Walter Lam directly. She introduced herself as Dr. Vasquez. The title felt like a coat she hadn't worn in years, stiff in the shoulders. She laid out what they had: the thruster trends, the theoretical mechanism from her 2070 dissertation, and Priya Chandrasekaran's experimental confirmation on second-gen Tesseral stock.
Walter listened. Asked sharp questions. He knew the material intimately, she could hear it in the precision of his objections, the way he reached for specific batch numbers and manufacturing tolerances without pausing.
"Dr. Chandrasekaran's samples are offcuts. Not panels under load. You can't extrapolate from a bench sample to an in-service panel without accounting for the active field environment, the gravitational coupling, the correction feedback loop."
Cory: "The grain boundaries don't know they're in a bench sample, Walter. They respond to thermal cycling. The profile Priya used matches the thermal history of a thirty-year-old building in a temperate maritime climate."
Walter: "It's an interesting result. And your dissertation mechanism is plausible. But you're asking me to accept that the fundamental substrate in six hundred buildings worldwide has a progressive aging defect that thirty years of monitoring have missed. You have a theoretical model, a bench experiment, and four noisy thruster curves that don't reach significance. That's a hypothesis. It's not evidence."
Cory: "What would convince you?"
Walter: "A quantitative model. A full computational simulation of the drift mechanism under real-world conditions, validated against field telemetry, with predictive capability. Show me the failure timeline. Show me when each panel crosses the threshold. Give me something that can't be attributed to sample artifacts or confirmation bias."
He wasn't being unreasonable. His objections were sound, his standards appropriate, his reading of the evidence defensible on its merits.
What he was wrong about was harder to name. There was a ten percent chance, maybe fifteen, that the trend was real. And if it was real, six hundred buildings were aging toward failure, including Antares V, a monumental station at fifteen kilometers altitude, four hundred thousand people, two hundred dollars an hour just for parking. Cory had never been. She'd had a coworker who saved for years just to visit for a single day. She'd seen the brochures. The zero-gravity water park, the holographic theater, restaurants serving food from neighboring planets. She'd thought about it as a cathedral she'd never afford to enter: wonder, quiet resignation, and nothing to be done about either. Now she was thinking about it differently.
The question was what you did when the evidence was uncertain and the consequences of inaction were catastrophic. That was a question about judgment, not statistics. And Walter had answered it by asking for more data.
She reached up and pulled the earpiece out, set it on the kitchen table, and sat in the quiet.
She'd known the conversation might go this way. Had prepared for it. What she hadn't prepared for was how it felt to hear someone dismantle her work, her dissertation, her prediction, Priya's confirmation, with careful, rigorous precision in the exact language of the world she'd left. In 2070, sitting in a lab at twenty-six, she had described exactly the mechanism that was now, eight years later, quietly shifting the grain boundaries in every anti-grav panel on the planet. She'd published it. Defended it. And then she got unlucky with the academic job market, so she walked onto a construction site and spent eight years building with the material she'd studied, hands cracked, paycheck to paycheck, twenty-five hundred dollars a month for a studio apartment in the shadow of a floating building she would never afford to live in. Now the prediction was coming due and a man from Tesseral had taken her apart in twelve minutes.
Something deep inside her cracked. She did not call Priya. She did not call Rua. She sat at the kitchen table and watched the Salinas Platform catch the late light through her window while tears ran silently down her face.
---
Priya called that evening. Cory answered because she always answered for Priya, but when she spoke, Priya heard it immediately, the difference between Cory-quiet-and-working and Cory-quiet-and-gone. Short answers. No engagement. Priya asked about the Walter call and Cory gave the summary in flat, efficient sentences. He wants a computational model. Validated against telemetry. Predictive capability. He was right to ask for it.
Priya was quiet for three full seconds.
Priya: "He was right to ask for more evidence. He was not right to stop looking."
Cory: "He's a materials scientist applying the appropriate standard of proof. I'd have done the same with this data eight years ago."
Priya: "No. You wouldn't. You'd have asked for more data and gone looking for it. That's what you did in 2070, you had a theoretical prediction and you designed an experiment. You didn't sit behind a standard of proof and wait for someone else to clear it. That's the difference."
Silence.
Priya: "Cory. The boundaries in my samples are migrating exactly the way your dissertation said they would. Exactly. Every sample. You called this eight years ago and the material is proving you right. That doesn't stop being true because a corporate gatekeeper is dragging his feet."
Cory: "It doesn't matter if I'm right. It matters if I can prove it to someone who can act."
Priya: "Then we prove it. Walter told you exactly what he needs. Fine. We know someone who can build that model. Probably the only person who can."
A long pause.
Cory: "Deshi hasn't responded to anyone in two months. Messages, calls, nothing."
Priya: "That's true. But Cory, Deshi will answer the door for you. You know that. The three of us sat in the grad office for years, and when Deshi went quiet back then, you were the one who knew how to be in the room without making it worse. I can't do that. I talk too much. That's not what Deshi needs."
Cory: "You're asking me to fly to Europe and knock on the door of someone I haven't spoken to in four years and ask them to come back to work."
Priya: "I'm asking you to go see a friend. The rest is between Deshi and the problem. But the problem can't get to Deshi without you."
Silence. Longer. Somewhere in Singapore, Priya was sitting in her lab at two in the morning, surrounded by sample slides that proved a material was failing, and she was spending this time not on the science but on the person. She did not experience this as a detour from the work. It was the work.
Priya: "Cory. Other people have studied this material at the atomic level — I'm one of them. But nobody else has done that and then spent eight years building with it. That combination doesn't exist anywhere else. Your hands felt something on that building that no instrument has detected. Rua believed you. I believed you. The samples believe you. Walter will believe you when we give him what he asked for. But I need you to still be here when that happens."
Cory closed her eyes. The kitchen was very quiet. Through the window, the Salinas Platform was a dark shape against the stars, its underside marked by the slow pulse of navigation lights.
Cory: "I'll go."
Priya: "Thank you."
Cory: "When did you get this good at managing people?"
Priya laughed. Short, surprised, warm. "I run a research group of eleven postdocs. It's basically like construction management, just with a little less aerogel."
---
Cory caught the Thursday morning flight from Monterey and landed in ZĂĽrich that evening, the descent rough in crosswinds over the lake. Three floating buildings, an office tower and two residential blocks, hung a kilometer over the old town center.
Deshi's apartment was in Wipkingen, on the slope above the Limmat. The building was old, stone, gravitationally conventional. A building that stayed where it was because it was heavy and nobody had paid to make it otherwise. Cory stood at the wooden front door for a full minute, watching her breath cloud in the cold, before pressing the buzzer.
No answer. She pressed again.
A long silence. Then, through the intercom, a voice she hadn't heard in four years. Quieter than she remembered.
Deshi: "Who is it?"
Cory: "It's Cory."
Another silence, longer.
Deshi: "Vasquez?"
Cory: "Yeah."
The lock clicked.
The stairwell smelled like old wood and espresso. Third floor. She climbed slowly, giving Deshi time, giving herself time.
The door was open a few inches. She pushed it and stepped in.
Dim. Curtains drawn. Not dirty. Deshi had always been tidy in a minimal, functional way. But stilled. The particular stillness of a space where nothing had been started in a while. A desk against the far wall held three monitors, all dark. Papers in careful stacks. A blanket on the couch that looked recently slept under. On the kitchen counter, a single mug, rinsed and inverted on a towel.
Deshi was standing by the window, holding the curtain open a few inches, looking out at nothing in particular. Thinner than Cory remembered. Same close-cropped hair, same watchful eyes, but the energy that had always burned behind them, restless, consuming, a restless, consuming focus that made everyone else in a room feel like they were standing still, was banked. Not gone. But far from the surface.
Deshi: "You flew to ZĂĽrich."
Cory: "I did."
Deshi: "That's either very kind or very alarming."
Cory: "Might be both."
Deshi let the curtain drop. Turned. Studied her for a moment with the particular attention Cory remembered — studying her as if modeling her from the inside out.
Deshi: "I've been getting messages. Priya. My department. I haven't." A pause. "I know I haven't been responding."
Cory: "You don't have to explain anything."
Deshi: "I'm not sure I can. That's part of the problem."
Cory nodded. She didn't move toward the couch. Didn't fill the silence. She knew this terrain, from inside her own version of it. Three months in a room like this after she left her postdoc. Different walls, same stillness. The feeling that the work you'd given your life to had become weightless, weightless, though without the elegance of EuGold — more the weightlessness of something detaching from meaning. She'd survived it by walking onto a construction site and putting her hands on physical things until the physical things started to matter. Deshi's way through would be different.
She sat on the floor by the desk, back against the wall. She bypassed the couch and the chair and sat on the floor by the desk, back against the wall. The floor was how they used to sit in the grad office when a problem got so big that chairs felt like pretending everything was manageable.
Cory: "I felt something wrong. With a building."
Deshi looked at her.
Cory: "Madison Heights. Office complex. Twenty-eight years old, second-gen panels. I was doing renovations, and I touched the wall before leaving. There was a lateral instability, very subtle. The kind of thing you feel before you understand."
She kept talking. Not fast, not urgently. Focusing on the essentials. The thruster curves. Rua's four buildings. The noisy data and Walter's rejection. Priya's samples. Her own dissertation, eight years old, coming due.
Deshi slid down the opposite wall and sat on the floor.
Deshi: "How fast does it drift?"
First real question. Cory answered carefully. Deshi asked another. Then another. Each one sharper than the last. Cory watched it happen, not like a light switching on, but like a system in standby beginning to draw power again. Deshi's eyes moved the way they moved when building a model internally: left, right, pausing, backtracking, locking on.
Deshi: "You're telling me the certification basis for every floating building in the world has an expiration date that nobody knows about."
Cory: "That's my concern."
Deshi: "Your extrapolation is linear."
Cory: "It's all I can do with a spreadsheet."
Deshi: "The real behavior won't be linear. Once drift exceeds the correction threshold, adjacent panels start coupling. Interfering. The whole system goes nonlinear." A pause. "It accelerates."
Cory: "That's why I need you."
Quiet for a long time. Outside, the light was fading. Cory watched the last sun slowly reflecting on the anti-grav panels of the floating buildings over Wipkingen.
Deshi: "I haven't opened my simulation environment in seven weeks."
Cory: "You have a new reason to."
Deshi: "I don't know if I can do this."
Cory: "You're the only person who can. And that's not flattery, it's the core constraint I'm working with."
Deshi almost smiled. It didn't quite arrive, but the architecture of it was there, the way scaffolding hints at what the finished structure will look like.
Deshi: "How much time do we have?"
Cory: "The drift is on the order of micro-degrees per year. We have time to do it right. I hope."
Deshi looked at the dark monitors. Then reached up, without standing, and reluctantly pressed the power button on the nearest machine. The screen bloomed to life, blue-white light falling across both of them on the floor.
Deshi: "Send me everything."
---
Deshi worked for six days.
Cory flew back to Monterey and worked in parallel with Priya, extending the experimental analysis, mapping drift profiles against building-specific thermal histories. They held a call every evening: Cory cross-legged on her apartment floor surrounded by printouts, Priya in her Singapore lab with sample slides racked in rows on every flat surface, Rua listening from her office with a coffee that was always fresh regardless of the hour. The picture sharpened day by day. Priya's cycling tests were unambiguous: grain boundaries rotated under sustained thermal load exactly as Cory's dissertation had predicted, and the self-correction mechanism held until it didn't. A quiet threshold with enormous implications.
Deshi surfaced on day six with a model that was, in Priya's words, "unreasonably elegant." It merged Cory's theoretical mechanism, Priya's experimental data, and the fleet-wide thermal exposure profiles that Tesseral published in its annual sustainability reports. It simulated drift behavior for every second-gen building in the world.
The results were worse than Cory's linear extrapolation.
Past year twenty-two, the drift accelerated nonlinearly. Buildings in high-thermal-gradient environments were furthest along the curve. The model projected that the first structures would exhaust their thruster compensation margin within two to four years. They would fall. Suddenly, silently, no warning.
Antares V was at the top of the list. Commissioned in 2053. Twenty-five years old. Highest sustained thermal gradient of any structure on the planet. If it fell, it would be the biggest tragedy in mankind's history.
Deshi's message was flat: Antares V has the least time. Two years, possibly less. The margin between correction failure and structural compromise is narrower than I expected. I have checked this. I wish the answer were different.
They assembled the report in forty-eight hours. Cory wrote the theoretical framework. Priya supplied the experimental validation. Deshi's model provided the projections. Rua, who understood what decision-makers needed and what they would dismiss, structured the document and cut every sentence that didn't carry weight.
Cory sent it to Walter Lam on a Monday morning: This is the standard of proof you asked for.
---
Walter read the report in his office at Tesseral headquarters in Delft. Rain on the windows. The coffee his assistant had left went cold on the desk.
He read it once in the morning. Read it again after lunch. Spent the afternoon pulling Tesseral's internal data and checking the model's predictions point by point against their own telemetry.
They held.
He went further. He ran Deshi's model against the full monitoring database, the one he'd told Rua would have caught a systemic trend. He looked at the data the way the model told him to: structured by age, thermal gradient, and batch-level grain density rather than building by building, where noise had always swamped the signal. The shape emerged. It had been there all along, in data he'd had access to every day for twelve years, disaggregated across six hundred buildings, invisible unless you asked exactly the right question in exactly the right way.
Evening settled over Delft. The rain had stopped. The office was quiet, everyone else gone home hours ago.
He took the small ceramic tile from his chest pocket and set it on the desk under the lamp. The iridescence shifted as it always did, light refracting along the quasi-crystalline axes in patterns that never quite repeated. He'd carried this piece of second-gen EuGold for twelve years. Knew this material the way a woodworker knows oak: by sight, by feel, by long acquaintance. Beautiful the way all materials are beautiful when their internal order is legible on their surface.
Twelve years of believing the material was sound. His belief was earned. He had tested it, studied it, defended it in regulatory reviews and certification audits. His belief was built on data, experience, and the intimate knowledge that comes from years of daily work with a single substance. But a construction worker had put her hand on a wall and felt what his instruments hadn't detected. A site manager had plotted four noisy curves and seen what his models had missed. A scientist had opened a 2070 dissertation and found the answer waiting. And now his own data was showing him the shape his own system had contained all along. Solid ground that suddenly felt unstable.
He picked up the handset on his desk and called Cory directly.
Walter: "Dr. Vasquez. The model validates against our internal telemetry. Point by point. It also holds against the full database when I structure the query by your colleague's variables. The signal was in our system. I should have found it."
A short silence. When Cory spoke, her voice was steady, but underneath it he could hear the relief. Not vindication. Just the relief of someone who has been carrying something heavy and who feels another pair of hands taking hold.
Walter: "When you called me with Dr. Chandrasekaran's bench data, I applied the right standard of evidence and the wrong standard of risk. The data was ambiguous. The potential consequence was not. I should have opened an investigation immediately and run the analysis I ran this afternoon. I didn't because I was certain the material was sound, and I let that certainty set the bar for action. That was my error, and I want to be precise about it."
Cory: "You're being precise about it now. And you know this material better than anyone alive. We need that."
He turned the tile under the lamp.
Walter: "What do you need from Tesseral?"
Cory: "Everything. Manufacturing specs, QC distributions, batch-level grain density data. Panel-by-panel thermal histories. Thirty years' worth." A pause. "And you. On the task force."
Walter: "My company's legal team is going to have opinions about that."
Cory: "I'm sure. Come anyway."
Walter: "Have your director invoice Tesseral for the workshop setup and crew time. Your outside specialists go on our contractor rate, backdated to when they started. And you get billed as a materials science consultant, Dr. Vasquez. Not a construction worker. Also backdated."
Cory was quiet for a moment.
Walter: "This isn't generosity. You identified a flaw in our product and built the proof. That's work Tesseral owed its clients. The least we can do is pay for it."
He put the tile back in his pocket.
Walter: "I'll be in Monterey on Wednesday."
That evening, Rua forwarded the invoice template. The line item read: Materials science consultant. The title had felt stiff when Cory first used it on the phone. It fit differently in writing. She filled in her hours — Altiva's consulting rate had kept her afloat, but this was something else, a number that reflected what she'd actually been doing — and sent it back.
---
Rua cleared the workshop floor adjacent to Altiva's main conference room, a concrete-floored space normally used for equipment staging, and had it set up as a temporary lab. She called it "requisitioning," though what she'd actually done involved relocating a CNC router, three crates of aerogel blankets, someone's bicycle, and one deeply confused intern. She'd also, through channels Cory decided not to ask about, sourced two additional workstations, a fume hood, and enough folding tables to furnish a small wedding. She had already sent Tesseral the first invoice before the workstations were plugged in.
Priya flew in from Singapore with two cases of archived EuGold samples and a portable thermal testing rig she'd disassembled and packed into checked luggage herself. She arrived Tuesday evening, inspected the workshop, and immediately began rearranging the bench layout without asking permission. Nobody stopped her. Deshi came from ZĂĽrich on Wednesday morning with three hard drives and a look that was cautious but present. The monitors at home, Cory suspected, had not gone dark again. Walter arrived from Delft an hour later carrying a box of panel shards, batch-level grain density records going back twenty years, and, in his chest pocket, a sample tile he didn't mention.
First time all four of them had been in the same room.
The conference space was utilitarian. Whiteboard walls, a view of the bay, a fire door propped open into the workshop where Priya's rig occupied two benches and Deshi's drives were already cloned to the workstations Rua had conjured. The coffee was bad. The aircon hummed. Cory felt, briefly and absurdly, like she was back in the grad office.
The first morning was tough. Walter's knowledge of the material was organized around assumptions the report had shattered. He kept reaching for explanations that no longer worked, stopping himself mid-sentence. A painful thing to watch in someone that precise. Cory's theoretical framework described the failure mechanism clearly; it did not describe a fix. Priya's data showed what was happening inside the panels; it couldn't tell them how to undo it. Deshi's model could predict when each building would fail. It could not prevent it.
They circled the same question for hours: could the drift be reversed?
Cory: "The grain boundaries have already moved. We can't push them back. The energy required would destroy the lattice."
Walter: "Then the panels are lost." Flat. His hand went to his chest pocket.
Priya, in the middle of a thought about something else, not quite looking at anyone: "The panels are different than they were. That's not the same as lost."
The room went still.
Priya stopped. Looked up. Then she was on her feet, pacing, three steps toward the whiteboard, pivot, three steps back. The way she moved when a thought was still forming and her body needed to keep up with it.
Priya: "Wait. We've been thinking about this backwards. All of us. We keep trying to return the panels to their original state. But what if we don't need to? The boundaries have moved. Fine. They've settled into a new alignment, that's where the material wants to be now. The problem isn't the new alignment itself. The problem is the correction mechanism is still calibrated to the original orientation. The panel is fighting itself. Holding a reference that doesn't match its own structure anymore. That mismatch. That's the whole thing. So what if we just… recalibrate to the panel as it actually is?"
Deshi's hands were already on the keyboard.
Walter was at the whiteboard before he'd finished standing, drawing the thermal curve of a correction cycle, annotated with parameters he knew by heart. Small, precise handwriting, filling the board fast.
Walter: "This might work. A thermal reset. We heat each panel through a controlled profile, specific ramp rate, hold temperature, controlled cooling, that re-locks the correction mechanism to the current grain boundary orientation. Like recalibrating a compass to a new north."
Priya: "Per panel. Each one has drifted differently. We'd need individual correction profiles."
Deshi, without looking up: "I can build those from simulations. Per-panel maps. Age, thermal history, batch-level grain density. Walter, I need your batch records for that."
Walter was already opening the box he'd brought from Delft. "Within the hour."
Cory watched his hand move across the whiteboard, fast and sure, material parameters flowing without hesitation. This was what twelve years of intimate knowledge were for.
They worked through the night. The coffee improved around midnight when Priya found a French press in a cabinet and commandeered it without discussion. Rua appeared at one in the morning with sandwiches and left again without comment. By dawn, Deshi had a prototype correction map for Madison Heights, all four thousand panels, each with an individually calculated thermal correction profile.
Priya proposed the rig design: a directed thermal source with a contact sensor for real-time feedback, correction profiles loaded from Deshi's maps. Simple, portable, operable in a service corridor by a trained technician.
Cory and Priya built the first prototype on the workshop benches, hand-soldering, cable-tying, swearing quietly at a stubborn temperature controller, while Walter and Deshi ran validation checks on the correction maps next door. Cory's hands moved with the practiced economy of eight years on construction sites: strip, crimp, test, move on.
Cory, tightening a mounting bracket: "This is too fiddly. It has to work in tight spaces, shift after shift. No specialized training beyond what a decent maintenance tech already has."
Walter, from the doorway: "And once it works, you need construction crews."
Cory: "We're going to need every crew Rua can find."
Priya loaded the first correction profile. The bench test went live an hour later, all four of them standing around the workbench, watching the data feed on Priya's laptop. She applied the thermal profile to one of her archived samples. The machine ran its curve. Priya held her palm near the rig and felt the heat. Precise, controlled, the exact profile Deshi's map prescribed.
On the screen, the grain boundaries responded. The drifted alignment absorbed the heat, relaxed, settled into its new reference state with a slow, clean certainty that looked, on the monitoring graph, like a long-held breath finally released.
Final readout: Angular lock stable. Self-correction nominal.
Priya carefully aligned the recalibrated tile to horizontal, held it in the air, and let go.
It floated. Motionless.
She trusted the analysis. But she also never felt settled until she could see it with her own eyes. She exhaled.
Walter assembled the third thermal recalibration prototype from Cory's design and handed it to Deshi, who ran it on panel shards from the Delft box. The finished device would need to be duplicated by the thousands, mass-produced, ruggedized, shipped to maintenance crews on six continents.
Walter took a breather. He pulled his personal EuGold tile from his chest pocket and turned it under the workshop light. The iridescence moved over it, complex, shifting, never repeating. He looked at it for a long time. Twelve years of love for a material doesn't break when you discover that it's flawed. It just changes shape.
Walter: "Forty thousand panels on Antares alone."
Cory: "And we know how to fix every one of them."
Through the conference room window, the late afternoon sun caught the underside of the Salinas Platform. Cory watched the faint shimmer of its panels, each one holding, for now, against a slow interior drift that nobody had seen coming because nobody had waited long enough to look.
The work ahead would be enormous. Months of coordination with station authorities and regulatory boards. Tens of thousands of individual panel treatments across six hundred buildings worldwide, starting with Antares and working down the risk list. Difficult, tedious, essential, the way all structural work is essential: invisible when it's done right, catastrophic when it isn't.
Walter, with a sigh, handed his personal tile to Deshi.
Walter: "Run this one too, please?"
Deshi took it carefully. Turned it once in the light. Set it on the bench beside the rig.
It was going to work out.
