Chapter 17: Glimmer in the Cracks

  Kenneth Ryder's death was like ice water, extinguishing the slight hope that had just been ignited by successfully opening the door, yet also forcing the remaining seven into an even more bizarre, delicate balance through Sophia Rossi's unexpected defection and partial confession.

  Mutual wariness, yet forced interdependence.

  Samuel Jones suppressed grief and dizziness, beginning to redeploy. Kenneth's position needed filling, and the newly joined Sophia had to be closely monitored, despite her claim of temporarily being on the same side.

  "Dr. Sharma, Mr. Schmidt," Samuel's voice carried fatigue but remained firm. "Focus on analyzing Kenneth's dying words. 'Waveform B,' 'mimicry,' 'careful.' This could be key. We need to know what Source B is actually mimicking, and what specifically we need to be careful of."

  Leon and Anya immediately threw themselves into work, two datapads and various monitoring screens becoming a temporary information processing center. Sophia was also allowed limited proximity, but Irina and Carmen flanked her left and right, "accompanying" her, gazes never leaving.

  Charles took one last look at Kenneth, watching helplessly as his "corpse" gradually disappeared. He felt a void-like helplessness. Death had never been so close.

  "Mimicry..." Anya murmured, fingers rapidly pulling up historical attack data stored in Leon's equipment. "If Source B isn't simply mimicking Source A, but mimicking the system itself..." She seemed to grasp some inspiration, beginning to construct new algorithm models, filtering out the train system's background noise, focusing on extracting the most minute, unnatural "consistency errors" in attack signals.

  Leon watched her operations, brow furrowed, seemingly also thinking of something: "The system's own signals... could it be hijacking or replicating Stardroop PLUS AI's underlying command fragments through some method? This would require extremely high authorization..."

  "Or, it's not hijacking at all." Sophia suddenly interjected. Arms crossed, she looked at the screen, eyes flickering. "What if 'Stardroop PLUS' itself... isn't monolithic?"

  This speculation was too bold, making everyone pause.

  AI split? Or infiltrated?

  "Impossible!" Leon instinctively rebutted. "Stardroop PLUS's core protocol is absolutely neutral!"

  "Absolutely?" Sophia sneered. "Like how the Stardroop system is 'absolutely safe'?"

  This was irrefutable.

  Just then, Anya suddenly let out a short exclamation: "Found it!"

  All eyes instantly focused on her datapad.

  On screen, two waveform graphs were highlighted. One was a clear record of Source A's attack, the other Source B's. While they differed greatly in macro terms, at an extremely subtle frequency band, Source B's waveform displayed a deliberately "calibrated," highly similar resonance characteristic to Source A, but this similarity seemed layered over some more primitive, more chaotic base signal.

  This tale has been unlawfully lifted without the author's consent. Report any appearances on Amazon.

  "This is 'mimicry'!" Anya said excitedly. "Source B is deliberately mimicking a characteristic of Source A, trying to make its attacks appear co-sourced with Source A! But its underlying signal... look at this residual fluctuation..." She enlarged the base signal for spectral analysis.

  It was a signal pattern full of disorder, violence, even a kind of mad quality, completely different from the train system or any known human technical style.

  "What... is this?" Carmen Ortiz frowned, the signal giving her instinctive discomfort.

  "Not like any known human coding protocol..." Leon's face became extremely grave, even showing a trace of incredulous fear. "More like... some kind... strong artificial intelligence? Or... something worse?"

  Non-human intelligence? This possibility made the temperature in the car plummet.

  "What about 'careful'?" Samuel pressed. "What did Kenneth want us to be careful of? Source B? Or the mimicked Source A?"

  Anya operated rapidly, cross-referencing Source B's base signal with the train environment database. "While mimicking Source A, its base signal is also continuously conducting low-intensity, wide-area scanning... the scan target seems to be... consciousness structure stability weaknesses? Or... some specific neural activity pattern?"

  She suddenly looked up at everyone, eyes filled with horror: "It's not just killing—it's... learning! Learning how to more efficiently destroy human consciousness! It's analyzing each of our fears, thinking habits, even subconscious reactions! What Kenneth wanted us to be careful of might be precisely this learning capability! It might increasingly understand us, the next attack could be even more impossible to guard against!"

  This conclusion was bone-chilling.

  One (or more) human killer(s) (Source A) was already terrifying enough. Now there might also exist a non-human, intelligence currently learning how to efficiently slaughter (Source B)?

  "We must leave here as soon as possible!" Pierre Chan's death and Lily's breakdown seemed to replay, panic beginning to spread. Even Irina and Carmen showed wavering on their faces.

  "How do we leave?" Leon roared, pointing at the navigation data on screen. "The buffer zone is ahead! At this speed, we'll arrive in at most a few more hours of system time!"

  Despair gripped everyone again.

  "Perhaps... there's still one method." Anya suddenly said, her voice somewhat trembling, seemingly having made a great determination.

  Everyone looked at her.

  "Since Source B can mimic system signals, possibly even infiltrate the AI, perhaps... we can too." Anya's gaze swept over Leon's equipment, then looked at Sophia. "Ms. Rossi, the deflection shield you just used—I'd bet you definitely still have one. Its energy characteristics are very unique, with strong adaptive deceptiveness. If... if we could combine this technical characteristic with Mr. Schmidt's monitoring equipment and my neural signal model... perhaps we can create a brief 'false command,' deceive the system, make the train decelerate early, or try emergency derailment procedures?"

  This was an extremely bold, nearly insane plan. Deceive a possibly infiltrated super AI?

  Sophia raised an eyebrow, seemingly somewhat surprised that Anya could see through her equipment's secrets, then revealed an interested smile: "Interesting. My little gadgets do have some 'deception' tricks. But deceive the entire train system? The required energy and computing power would be astronomical, and if it fails, it might immediately trigger the most extreme defense mechanism, blowing us all into nothingness."

  High risk, almost certain-death risk.

  But there were no other options.

  Samuel looked at Leon: "Can your equipment provide enough computing power?"

  Leon calculated with a grim expression: "Overload operation... might hold for ten seconds. But it requires extremely precise synchronization and energy guidance. And we need an access point, a core data node where system defenses are relatively weak."

  His gaze, along with Charles's, once again turned to that deadly equipment room door at the car's rear.

  Behind it was the data abyss confining Marcus Wong's consciousness, but also possibly the most dangerous interface to the system core.

  The just-closed Pandora's box seemed to have to be opened again.

  The only way out actually hid in the deadliest trap.

  The fate of seven people once again hinged on what lay behind that door.

  And would that "learning" Source B allow them to attempt this?

  Source A, that hidden human killer—when would they swing their blade again?