Sera Vandermeer talked for forty-seven minutes without stopping.
This was not, Rivan understood, because she lacked the discipline to be concise. It was because she had been carrying this particular weight alone for fourteen months and had made a decision somewhere between the Telegram messages and the flight from Singapore that she was going to trust this conversation, and having made that decision, she was going to honour it completely.
He did not interrupt. He listened the way he had trained himself to listen during twelve years in markets not passively, not with the half-attention of someone composing their response while the other person was still speaking, but with the full, committed presence of someone who understood that information presented in sequence contains a structure, and that the structure is often more revealing than the content.
What emerged over forty-seven minutes was this:
Sera had joined her firm a Singapore-based quant shop called Meridian Analytics in early 2018, straight from a master's degree in financial mathematics at Leiden University in the Netherlands. Her father was Indonesian, her mother Dutch, and she had grown up between the two countries in the way that children of dual-national families often do fluent in both languages, fully at home in neither, with the particular quality of perceptiveness that comes from spending a childhood reading rooms where you are always slightly the outsider.
She had noticed the Soerjo Capital anomaly in January 2018, during her third week at Meridian, while running a routine cross-market correlation analysis for a report on Southeast Asian institutional crypto entry. A small discrepancy in the data a cluster of wallet movements that did not fit the behavioral profile of any known institutional actor had caught her attention the way small discrepancies always caught the attention of people trained to find them.
She had flagged it to her supervisor.
Her supervisor had told her it was noise and to move on.
She had moved on, officially. Unofficially, she had begun building a separate dataset on her personal hardware, working on it in the early mornings and late evenings, the way people work on things they are not supposed to be working on.
"Your supervisor told you it was noise," Rivan said, when she reached this point. "Did you believe him?"
"No."
"Did you think he was wrong, or did you think he was lying?"
Sera looked at him with the careful attention of someone encountering a question they had asked themselves and not fully answered. "At the time, I thought he was wrong. Analysts dismiss anomalies all the time it's a standard error, especially in crypto where the signal-to-noise ratio is genuinely terrible." She paused. "Now I think he may have been lying. I can't prove it. But the Meridian Analytics client list has three entities on it that, based on my current dataset, are almost certainly connected to the Soerjo Capital network."
"You're still at Meridian," Rivan said.
"Yes."
"And they don't know about your personal dataset."
"I've been careful." She said it with the flat confidence of someone who had thought carefully about operational security and implemented it methodically. "Separate hardware. No network connection during active research. Physical notes only for anything sensitive. I've been treating it like a personal project on financial contagion risk in Southeast Asian markets which, technically, it is."
"How long can you maintain that cover?"
"Until I publish something or until someone pays close enough attention to my personal schedule." She looked at him directly. "Which is why I'm here. I've taken this as far as I can alone."
She showed him the dataset.
Not all of it she had printed selected sections, the key nodes and connections, with enough context to be meaningful and enough redaction to be manageable. He spread the pages across the coffee place table and read through them with the systematic attention of a man who had been building a similar picture from a different angle, and felt the specific, uncomfortable sensation of watching two incomplete maps resolve into something more complete.
The Singapore-side data was cleaner than his Jakarta-side view. Sera had access to cross-border capital flow analysis that showed the full circuit money entering crypto markets from Soerjo Capital's investment vehicles in Jakarta, moving through Singapore intermediaries, accumulating in wallet structures that were technically independent but behaviorally correlated. The circuit was elegant. It was also, he noted, exactly the kind of structure that would have been invisible to anyone looking at only one end of it.
"The Singapore intermediaries," he said, pointing to a cluster of nodes in the center of her diagram. "These three entities. Do you have names?"
"Two of them, yes. The third is behind a layer of nominee structures I haven't been able to penetrate." She tapped the unnamed node. "This is the most active of the three. It's been receiving the largest volume of Soerjo Capital flows, and it's been distributing them into crypto positions with a timing precision that is not humanly possible without either extraordinary luck or a predictive system."
Rivan looked at the unnamed node for a long moment.
"A predictive system," he said.
"Something that's reading market microstructure in real time and identifying optimal entry points with a consistency that no standard algorithm should produce." She looked at him. "I know how that sounds."
"It doesn't sound strange to me," he said carefully.
"No. I didn't think it would." She watched him with the particular attention she had been applying since the moment she sat down not suspicious, but thoroughly observant. "You have an edge that isn't coming from standard analysis. I've been trying to determine what it is since I first noticed the @g_nara account. The timing of your online presence relative to market events is too precise for alert-based monitoring. It's more like you know before the alert would trigger."
Rivan said nothing.
"I'm not asking you to explain it," she said. "I'm telling you that I've noticed it, and that it's one of the reasons I'm sitting here. Whatever your edge is it's real, and it's significant, and it's consistent with what I'd expect from someone who had access to the same kind of predictive capability that the unnamed Singapore node appears to have."
The implication hung in the air between them, fully formed and unspoken: are you connected to the network you claim to be investigating?
"I'm not connected to Soerjo Capital," Rivan said.
"I know," she said.
"How?"
"Because if you were, you wouldn't be watching from a secondary account in a mid-tier Telegram group. You'd have no reason to." She paused. "Also because your accumulation patterns are building capital in a way that's consistent with someone operating independently on a limited budget, not with someone who has access to institutional resources. I've been tracking the on-chain footprints that are probably yours for six weeks. The position sizes are too small for institutional involvement."
Rivan looked at her.
She had been tracking his on-chain footprints for six weeks. Before she had messaged him. Before she had arrived in Jakarta. She had found him not his username, but his actual market activity and spent six weeks building a profile before making contact.
He was not certain whether to be impressed or concerned. He settled on both.
"You're thorough," he said.
"I try to be." Something in her expression shifted not quite amusement, but adjacent to it. "You're about to ask whether that level of surveillance makes me a threat. The answer is that the same skills that let me find your footprints let me find the Soerjo Capital footprints. Both processes look the same from the inside. What matters is what I do with the information."
"And what are you going to do with it?"
"That," she said, "depends on what you tell me about the other three people."
He told her about Laras first the thesis, the correlation dataset, the four months of work. Sera listened without interrupting, made two notes, asked one clarifying question about the methodology Laras had used.
He told her about Dr. Mira Salim second the four years of research, the network diagram, the three researchers who had disappeared.
This part Sera received differently her expression tightened in a way that suggested the information was landing somewhere specific, connecting to something she already knew.
"The Seoul researcher," she said. "2018."
"You know about that."
"I know about a paper that was retracted, yes. It was circulating in certain channels before it disappeared I read it." She was quiet for a moment. "The methodology was sound. The conclusions were conservative, actually more conservative than the data supported. Whoever retracted it didn't do it because it was wrong." She looked at him. "Who is the third person? The one with four years of data?"
"Not yet," Rivan said.
She accepted this with a small nod the acknowledgment of someone who understood that trust was built in increments and that asking for the full amount upfront was a category error.
"The unnamed Singapore node," Rivan said.
"I want to know more about it."
"I thought you would." She reached into her folder and produced a single page separate from the dataset, printed on different paper, which suggested she had prepared it specifically for this moment. "This is everything I have on it. Which is less than I'd like, but more than nothing."
He took the page.
At the top, she had written: Entity classification: Unknown. Behavioral profile: Analyst/Coordinator type. Geographic anchor: Singapore, probable Raffles Place district. Active since: Q3 2017.
And at the bottom, in a section labeled Possible identifiers:
Username observed in three separate dark web financial forums (2017-2018): @v_salim_research
Rivan looked at the username for a very long time.
@v_salim_research.
The Salim could be coincidence. It was not an uncommon name. The research designation could be coincidence too many forum usernames referenced professional backgrounds.
He thought about Dr. Mira Salim, visiting researcher, Economics faculty, University of Indonesia. Who had been building her dataset since 2015. Who had contacted him at 2:34 AM through an anonymous channel with the specific details of a warung near his university. Who had known about The Curator before he had. Who had told him, with the flat certainty of a scientist confirming established data, that Prof. Handoko Wirawan was the man at the top of the network.
He thought about Laras Andini, who had been pointed toward him by someone named Bimo he had never been able to locate.
He thought about Sera Vandermeer, who had arrived in Jakarta two years and nine months before she was supposed to exist in his timeline.
He thought about the GHOST Protocol's recursive loop warning: Cannot assess NULL_SIGNAL because NULL_SIGNAL has partial knowledge of GHOST Protocol.
And he thought about the question Dr. Salim had left him with, under the blue tarp, on a Wednesday morning in March:
Who sent you back and what are they not telling you?
"I need to make a call," he said.
Sera looked at him. "Right now?"
"Yes." He was already reaching for his phone. "It won't take long."
He opened the contact channel Dr. Salim had given him after the warung meeting the one she had said to use only for significant developments. He typed a message in the careful, non-specific language they had agreed on for sensitive communication:
"The Singapore researcher I mentioned. I've seen a username. @v_salim_research. Can you confirm this means nothing to you?"
He put the phone face-down on the table.
Looked at Sera, who was watching him with the patient, thorough attention she brought to everything.
"While we wait," he said, "tell me about the unnamed node's trading behavior. Specifically has it ever made a wrong call?"
"Once," she said. "March 2018. A significant ETH position that was entered at the wrong time and exited at a loss three weeks later." She paused. "What's interesting is the timing of the error. It was the same week that the Seoul research paper went dark."
Rivan looked at her.
"The entity made a mistake," Sera said quietly, "in the same week that someone who had found part of the same thread was silenced. As though something that had been providing the coordination layer with accurate data had been disrupted. Temporarily."
His phone buzzed.
He turned it over.
Dr. Salim's reply consisted of four words:
"We need to meet."