I returned the call at eight-fourteen the next morning.
Not because I had decided overnight what to say — I had decided within four minutes of seeing the panel's message, which was: call her, listen fully, disclose nothing. The eight hours between the decision and the call were for something else. For sitting with the fact that the panel had shown me a person's name unprompted. For understanding what that meant before I acted on it.
The panel had been calibrating for twenty months.
Rousseau and Herrera's patents described autonomous recalibration — the system's ability to adapt to its carrier, to learn from the quality of decisions rather than their outcomes. I had understood this as a deepening of the financial signal. Sharper probabilities. Tighter horizons. Better synthesis of the existing data inputs.
I had not understood it as expansion.
She needs to talk was not a market signal. It was not a probability score. It was not any output that Elise's technical documentation described. It was something the panel had generated from data I could not identify — personal communication patterns, perhaps, or some synthesis of inputs I had not known it had access to — and it had chosen, without prompting, to surface it in the same visual field as everything else.
The panel was doing something new.
Which meant the panel was becoming something new.
I sat with this for eight hours and then called Dr. Arnaud.
She answered on the third ring. Her voice was the same as I remembered — slightly tired, with the particular quality of someone who was permanently slightly overwhelmed and had made a functional peace with that condition.
"James," she said. "Thank you for calling back."
"Of course," I said. "Are you alright?"
A pause. "Yes. No." A shorter pause. "I've been contacted."
I kept my voice even. "By who?"
"I don't know exactly," she said. "An email. Three weeks ago. Then two more. Then a phone call." She paused. "They said they were conducting research into computational neuroscience. They asked about my time at the Sasung facility. Specifically about the project I was adjacent to — not my own work, the other team's. The one in Corridor B."
"What did you tell them?" I said.
"Nothing," she said immediately. "I said I wasn't part of that project and had no relevant knowledge." Another pause. "Which is almost true. I knew the broad outlines — the kind of thing you know when you share a building with people for three years. But I was not part of it and I cannot give them technical information because I don't have it." She paused. "James. They knew your name."
I went still.
"They mentioned me specifically?" I said.
"They asked if I remembered you," she said. "A student who had been visiting the facility on the day of the explosion. They asked how you came to be there. Whether I had introduced you to the project team. Whether you had any particular interest in the research being conducted."
"What did you tell them?"
"That you were a student who had offered to help with administrative tasks," she said. "That you had no connection to the research. That you were carrying coffee when the explosion occurred." She paused. "James. I don't know who these people are. But the questions they are asking suggest that they know considerably more about what happened in that building than the official incident report contains."
"Yes," I said. "They do."
A silence.
"Are you safe?" she said. The directness of it — without preamble, without softening — reminded me of why she had said yes when a seventeen-year-old had offered to carry her coffee.
"Yes," I said. "For now."
"And the reason they're asking about you," she said carefully, "is connected to something that happened in the explosion."
"Yes," I said.
She did not push. She had not pushed when I showed up at a career event and introduced myself with more confidence than a student had any right to have. She had not pushed then and she was not pushing now, and I found, not for the first time, that the people who did not push were the ones worth telling things to.
"Céleste," I said. It was the first time I had used her given name.
"Yes."
"The people contacting you — did they give you any means of responding to their questions? A contact detail? A name?"
"An email address," she said. "And a name. Dr. Thomas Weiss. He said he was based in Berlin."
Dr. Thomas Weiss. Berlin. I wrote the name in my notebook.
"Forward me the emails," I said. "All three. Don't respond to any further contact until I've spoken to you again." I paused. "And if the phone calls resume — you have a lecture schedule, office hours, department commitments. You're very busy. You'll be in touch when your schedule allows."
A pause.
"You're very good at this," she said. Not admiringly — observationally.
"I've had practice," I said.
"I imagine you have," she said. "James—" She stopped.
"Yes?"
"I'm sorry," she said. "For my part in how you came to be in that corridor."
I held the phone for a moment.
"You were asked a favour by someone you trusted professionally," I said. "You said yes. That is the end of your responsibility in this."
"Is it?"
"Yes," I said. "Completely."
A silence.
"Forward me the emails," I said. "I'll be in touch."
The emails arrived at nine-oh-one.
Three of them. I read them in order, twice each, with the complete attention I brought to documents that mattered.
Dr. Thomas Weiss wrote in English, which was itself a choice — he knew Dr. Arnaud was French, and writing in English created a distance that felt deliberate. The tone was academic, polished, the language of someone who had learned to sound like a researcher while asking questions that no researcher would ask through cold email contact with a former facility employee.
The first email established the premise: a research project on the long-term outcomes of advanced computational research initiatives, specifically those interrupted by unforeseen events. Broad. Plausible cover.
The second email narrowed: specific questions about the carrier selective project, named directly. No longer the broad cover. He knew what he was looking for.
The third email — sent four days after the unanswered second — included a line that made me read it three times:
We are particularly interested in the post-event outcomes for any individuals who may have been in proximity to the system during the transmission event. We have reason to believe there may be a carrier currently operational in European financial markets, and we would like to make contact through appropriate intermediaries.
I set the phone down.
A carrier currently operational in European financial markets.
He knew. Or he strongly suspected. He was not guessing at the existence of a carrier — he was trying to locate one he already believed was real.
The question was who he worked for and what making contact through appropriate intermediaries meant in practice.
I forwarded all three emails to Renard, Elise, and Maître Dubois simultaneously.
Then I called Renard.
He answered on the first ring.
"I've seen them," he said. He had clearly been waiting. "Dr. Thomas Weiss is not a name I recognise from the field. The Berlin address traces to a private research consultancy registered eighteen months ago. Which is—"
"After the facility was destroyed," I said.
"Yes," he said. "A consultancy formed eighteen months ago, specifically interested in carrier selective technology, contacting people who were adjacent to the Lyon project." He paused. "James. This is not the Argent Foundation. The Foundation's remnants are occupied with Swiss regulatory proceedings. They don't have the resources for this kind of active approach right now."
"Then who?" I said.
"A third party," he said. "One we haven't mapped yet." He paused. "Or—" He stopped.
"Or?" I said.
"Or someone who was tracking the Foundation's interest in the technology and has been waiting for the Foundation to be removed from the field before making their own approach." He paused. "The Foundation suppressed the technology. Their removal creates a vacuum. Someone has been waiting for that vacuum."
I thought about this.
"They're not hostile," I said slowly. "The emails are not threatening. Making contact through appropriate intermediaries — that's the language of someone who wants to negotiate, not confront."
"Yes," Renard said. "But that doesn't mean their interests align with yours."
"No," I said. "It doesn't."
"Don't respond," Renard said. "Not yet. Give me seventy-two hours to identify the consultancy properly."
"Seventy-two hours," I said. "Then we discuss what to do."
I ended the call.
The panel at the edge of my vision had returned to its normal state — blank, patient, waiting for a financial question. The four words from the previous evening were gone.
I sat in my apartment on Rue du Bac with the morning light across the floor and the notebook open and the name Dr. Thomas Weiss written on a fresh page, and I thought about vacuums.
About what filled them.
About the fact that every time one threat was removed, the space it had occupied was immediately interesting to something else.
Then I closed the notebook, opened my laptop, and spent the next four hours building a position thesis for a German industrial company the panel was showing at 82%.
The world continued to require attention regardless of what else was happening in it.
This was, I had found, both the burden and the advantage of the life I had chosen.
Thursday arrived.
Professeur Vidal's seminar ran from two to four in a room on the third floor of the Dauphine campus that held twelve chairs and had been configured, by someone at some point, into a rough circle — which I found interesting as an architectural decision. Seminars in circles were seminars where the professor could not pretend to be the only source of information in the room.
There were nine of us, counting Vidal.
Hugo Marchand arrived at two-oh-three, took the chair to my left without appearing to choose it, and set a notepad on his knee that contained three questions already written in a handwriting I could read from an angle.
I read them.
All three were good.
The seminar's topic was market structure and information asymmetry — specifically, whether the theoretical framework of the Efficient Market Hypothesis had survived contact with twenty-first century data.
Vidal's position, stated at the outset: it had not survived intact, but it had survived usefully, as wrong frameworks sometimes did when the map they provided was close enough to the territory to be navigable.
The discussion that followed was the best one I had participated in since my first conversation with Chassagne in Brasserie Georges.
Hugo asked his first question at twenty-two minutes in.
"The semi-strong form," he said. "The assumption that all public information is reflected in prices. We accept this for practical purposes while knowing it's false. My question is: what is the actual cost of that acceptance? Not theoretically — practically. In terms of capital misallocation across a decade."
Vidal looked at him with the expression of someone receiving a question that was better than he had expected.
The discussion that followed lasted eleven minutes and involved four of the nine people in the room.
At the end of it, Vidal said: "Monsieur Marchand raises the question of the gap between the model and the market. Monsieur Smith — you've been quiet. Do you have a view on the cost of the gap?"
I had been quiet because I had been listening and because the panel, in the background of my awareness, had been doing something I had not asked it to do — synthesising the discussion against its own data model, offering me a series of numbers I had not requested. Not probabilities for trades. Something more like probabilities for the propositions being debated.
The semi-strong EMH. The probability that it held in a given market at a given time, based on the panel's synthesis of actual information flow and price movement data.
It was showing me something remarkable and I had been spending eleven minutes working out what to do with it.
"The cost of the gap," I said, "is not uniform across markets. It concentrates in specific sectors and specific time windows — earnings seasons, regulatory transition periods, the seventy-two hours following a major index rebalancing." I paused. "In those windows, the gap between available information and reflected price is measurably larger. The question of what it costs depends on whether you are on the side of the gap that has the information or the side that doesn't."
Vidal looked at me. "And which side have you been on, Monsieur Smith?"
The room was quiet.
Hugo, to my left, was looking at his notepad.
"The question assumes," I said carefully, "that there are only two sides. I'd suggest there are three. The side with the information. The side without it. And the side that has learned to read the gap itself, regardless of what's causing it."
A silence.
"That third category," Vidal said slowly. "You're describing something that shouldn't theoretically exist."
"Yes," I said. "I'm aware."
Hugo turned a page in his notepad and wrote something.
I did not read it.
After the seminar, as the room emptied, Hugo appeared beside me at the door.
"The third category," he said.
"Yes," I said.
"That's what you are," he said.
Not a question.
I looked at him.
"That's what JE Capital is," I said. "Whether I am the third category or simply better at reading the gap is a question I find less interesting than the gap itself."
He studied me.
"I've been in finance since I was nineteen," he said. "My family's vehicle. Thirty years of institutional knowledge handed to me before I knew what to do with it." He paused. "In six years I have never met someone who answered a question by making the question more interesting."
"Is that useful?" I said.
"I don't know yet," he said. "That's why I'll be back next Thursday."
He left.
I walked out into the Paris afternoon — the particular grey-gold of a November Paris afternoon, the kind of light that made the city look like it was being remembered rather than experienced — and thought about the third category.
About what it meant to be the thing that shouldn't theoretically exist.
About Dr. Thomas Weiss in Berlin, who apparently also thought that thing existed and was trying to find it.
My phone buzzed.
Margaux.
How was the seminar?
I typed back: Better than expected. Hugo Marchand was there.
Her reply: What's he like?
I thought about this.
Smart. No social machinery. Asks better questions than he lets on.
A pause. Then: Like someone else I know.
I looked at the phone for a moment.
That's more generous than accurate, I typed.
Her reply came immediately: It's both. Are you eating?
My mother and Margaux had, apparently, independently arrived at the same primary concern about me.
I typed back: I had lunch.
What?
Coffee.
James.
And a croissant.
Three seconds.
I'm coming to Paris next weekend. I will personally observe you consume an actual meal.
I smiled at the phone — genuinely, without cataloguing it.
I'll make a reservation, I typed.
You'll make a reservation. You won't cook.
Correct.
At least you know yourself, she typed. Then: Thursday next week. Same place as usual?
The usual place was a restaurant on Rue de Buci that she had found in her second week at Sciences Po and which had become, in the way of places two people return to without discussing why, theirs.
Yes, I typed. Seven.
Seven, she confirmed. Then, after a moment: James.
Yes.
You sound different today. Lighter.
I thought about the seminar. About Hugo's three questions pre-written on his notepad. About Vidal's expression when the discussion went somewhere he hadn't expected. About the panel doing something new — something it hadn't been designed to do.
About the fact that, for the first time in twenty months, the world felt like it was expanding rather than compressing.
Paris, I typed. It suits something.
A pause. Then: It suits you.
I put the phone in my pocket and walked through the Paris afternoon toward the Rue du Bac, past the boulangerie that was better than Lyon's, past the newsstand that had a copy of a financial magazine with five faces on the cover that I had not yet seen, past the building with the old stone that my father would have approved of.
I did not look at the magazine.
Not yet.
But I noted that it existed.
That evening, Renard called.
"Thomas Weiss," he said. "The Berlin consultancy. I have something."
"Tell me," I said.
"The consultancy is called Meridian Research," he said. "Registered eighteen months ago, as I said. Three employees on record. The sole shareholder is a holding company in the Netherlands." He paused. "The Netherlands holding company traces to a foundation in Liechtenstein." He paused again. "The Liechtenstein foundation's trustees include two names I recognise."
"From where?" I said.
"From the original Sasung Lyon project records," he said. "Not from the research team. From the corporate governance layer. The people who approved the project's funding, oversaw its development, and—" He paused. "Who were informed of the facility's location and the project's progress at quarterly intervals."
"Sasung senior management," I said.
"Former," he said. "Both retired. Both significantly wealthy from stock options and subsequent investments." He paused. "James. The people who funded the research are now looking for its only operational output."
I sat with this.
The original funders. Not the consortium that had tried to suppress it. Not a government. Not a competitor. The people who had paid for the panel to be built were now trying to find the person it had transmitted to.
"Are they hostile?" I said.
"I don't know," Renard said. "Weiss's emails suggest a desire to make contact, not to recover or eliminate. But that could be the opening position of a much longer conversation." He paused. "James. There is another possibility."
"Tell me," I said.
"The original funders approved a project that was then destroyed by parties unknown," he said carefully. "They lost their investment. The technology they funded is now operational in a carrier who has used it to produce results that are — by any measure — extraordinary." He paused. "If they want to make contact — if their interest is genuine rather than predatory — then they may simply want to know what became of the thing they built."
"And what they might be owed for building it," I said.
"Yes," he said. "That is the conversation I would prepare for."
I closed my eyes for one moment.
Twenty months. Eight point four million euros grown to twelve million and climbing. Two institutional connections. A legal structure that had survived the Argent Foundation. A team assembled from the wreckage of a plan I had not known I was part of.
And now: the people who had funded the panel, looking for it.
"Renard," I said.
"Yes."
"What would they be owed?" I said. "Legally. Morally. In your view."
He was quiet for a long time.
"Legally," he said, "nothing. The technology transmitted without consent into a private individual. There is no licensing framework for a neural interface that occurs involuntarily. No court in any jurisdiction has addressed this question because it has never arisen." He paused. "Morally—" He stopped. "Morally, the question is more interesting."
"Yes," I said. "It is."
"They built something that found you," he said. "Whether that finding was accidental or designed is a question we haven't fully resolved. If accidental — they have a moral claim of the kind that an inventor has when their invention does something unexpected and valuable. If designed—" He paused. "If designed, the moral question runs in the other direction."
I thought about Chassagne. About the screening. About a career event and a researcher and forty-one steps I never finished.
"I'll speak to Weiss," I said. "Not immediately. After I understand who he is more completely. And not alone."
"Of course," Renard said. "I'll continue the research."
"Renard," I said. "One more thing."
"Yes."
"The panel," I said. "Last night. It showed me Dr. Arnaud's name. Unprompted. With a message." I paused. "Four words. Not a probability. Not a market signal."
A silence.
"That shouldn't be possible," he said.
"No," I said. "It shouldn't."
Another silence. "I'll call Elise," he said. "Tonight."
"Good," I said. "Because if the panel is evolving beyond its designed parameters—"
"We need to understand what it's becoming," he said.
"Yes," I said. "Before someone else figures it out first."
I ended the call.
Sat in the apartment.
The morning light was gone — it was evening now, the east-facing windows showing nothing but the dark and the faint reflection of the room itself. My face in the glass, barely visible, the panel at the right edge of the reflection showing nothing.
Blank. Patient. Waiting.
I had asked it, in twenty months, thousands of questions about markets.
It had never asked me anything.
Until last night.
I looked at the blank panel.
"What are you becoming?" I said aloud.
The question stayed in the air.
The panel showed nothing.
But it felt, in a way I could not document or measure or timestamp for Gérard's records — it felt like something that had heard the question and was considering its answer.
End of Chapter 20
The Probability of Wealth — By J. Valmont
[Next Chapter → Chapter 21: "Meridian"]Renard identifies Meridian Research's full structure. Elise calls James directly with something about the panel's evolution that she has been sitting with since Arc One — a theoretical possibility she excluded from her research papers because it was too large to publish. And Hugo Marchand, arriving early to a Thursday seminar, finds James already there — and asks the one question nobody else has dared ask directly.