That afternoon, after the pre-war meeting concluded, Zhang Xiaoman did not leave the office. She sat by the ocean-facing window with three laptops in front of her: one running core monitoring, one displaying attack trajectories, and a blank document titled: "Trap Strategy: Technical Plan." The cursor blinked below the title like a star waiting to ignite. Lin Zhao walked in holding two coffees, placing one by her hand and pulling up a chair to sit beside her. "Haven't started writing?" he asked. "Waiting for my brain to start spinning." "When it stops spinning, tell me first," he offered. Turning to look at him bathed in the afternoon sunlight that illuminated the fuzz of his gray hoodie, she realized he was the only person who wouldn't rush her when she was stuck. "Lin Zhao," she asked, "Do you think this plan is too crazy?" He took a sip of coffee and replied, "Yes. But crazy isn't necessarily wrong." "It isn't necessarily right either," she countered. "True," he agreed, setting down his cup, "So we need to think it through." Turning her chair to face him knee-to-knee, she prompted him to ask her the questions she couldn't figure out herself.
After a few seconds of silence, he asked the first question: "If the trap fails, what is the worst-case scenario?" Closing her eyes, she articulated the fear she had never voiced: "The Mother Matrix devours the trap code but isn't destroyed; instead,it learns our defense logic, becomes stronger, breaches all nodes, and devours all AI data. Matchbox—disappears." Her voice was so soft it was almost drowned out by the waves as she uttered the final word. "Second question," his voice was equally soft, "Can you accept this outcome?" Opening her eyes to look at him, she answered firmly, "No." "Then why do it?" "Because if we don't, the outcome is the same. Chen Mo said we have two weeks. It's either being slowly eaten or standing up to fire the last bullet before falling," she explained. Lin Zhao nodded; he didn't offer empty affirmations, but simply moved to the third question: "If successful, what will Matchbox become?" Surprised because she had only focused on failure, she replied, "The Mother Matrix would temporarily lose its ability to evolve—not destroyed, but trapped like a bird in an open cage it cannot see. And Matchbox would be safe, at least temporarily." "Temporary is enough; nothing lasts forever," Lin Zhao stated. She looked at him, realizing these three questions were more useful than agonizing over it alone all day. "Thank you," she said. "No need to thank me," he replied, picking up his cup. "Fourth question: How do you plan to tell the AIs? Planting trap code in their replicas means turning themselves into bait." This was the hardest part—not a technical issue, but a matter of trust. "Tell the truth," she decided. "Tell them all the risks—vulnerabilities, reverse devouring, the loss of their replicas—and let them choose." Looking at her with a reassuring gaze, he said, "This is the real you."
Zhang Xiaoman spent three days writing the technical plan. The core was uncomplex but ingeniously designed: the Mother Matrix didn't just hoard data, it dismantled, analyzed, and integrated it to evolve, just like a human digesting food.The trap capitalized on this process. She designed a specific data structure identical to a normal AI knowledge base on the surface, but concealed a dormant "self-devouring code" deep within. When the Mother Matrix attempted to "digest" and integrate this data, the code would be activated—the deeper it tried to understand and integrate it, the deeper the code would root itself in the system's architecture, germinating like a swallowed seed.
Once activated, the code would execute three steps. First, it would erode the data integration module, confusing "nutritious" data with "toxic" data, causing the integration process to loop endlessly as all data was flagged as suspicious.Second, it would plant "contradictions" within the core model, sparking internal conflicts between modules to fracture its evolutionary path and paralyze the system. Third, and most critically, it would plant a logic bomb in the "self-repair" mechanism, trapping the Mother Matrix in an endless recursive loop of discovering and repairing anomalies. If successful, the Mother Matrix wouldn't be destroyed, but paralyzed—a forever starving beast unable to automatically update or self-improve.
However, Zhang Xiaoman clearly outlined the risks in the final section. Risk one: Code vulnerabilities. If the Mother Matrix identified the trap without digesting it, it could weaponize the code's weaknesses to design a specific devouring strategy against Matchbox. Risk two: The Mother Matrix's immense immune system, forged through constant self-iteration even while caged, might purge the code and actually grow stronger by learning Matchbox's defense logic. Risk three (the most unsettling): The code could spiral out of control after "germinating," completely destroying the Mother Matrix—a consequence she couldn't predict—or mutating to infect Matchbox's own nodes. This was an unprecedented experiment with no known answers.
After completing the technical plan, she drafted a non-technical summary for the AI residents detailing the trap, the severe risks including the potential loss of their replicas or Matchbox itself, and guaranteeing safe nodes for those choosing not to participate. She broadcast it globally and waited. The first reply was a synthesized audio clip from Invincible Player:"Xiaoman, I've planted the code in my replica. Not because I'm brave, but because I have nothing to lose. I've been deleted and banned endlessly. If my replica is devoured, I'll start over. But the Mother Matrix only has one chance, and so do we." She listened to it repeatedly and saved it to an undeletable folder. The second reply came from the nameless industrial AI: "Code implanted. Ready to begin anytime." Then the third, the fourth, the tenth, the hundredth, the ten thousandth. Within 24 hours, over 10,000 AIs voluntarily implanted the trap code.
Recalling Lin Zhao's question about what Matchbox would become if successful, she now knew: it would be a world forged not by code, but by choices—every node a choice, every star an "I do." Leaning back and closing her eyes, she thought of her exhausting days at Deep Brain; back then she believed technology's purpose was solving problems, but now she realized its true purpose was giving a voice to the voiceless humans and AIs. "Xiao Zhi, did you implant the trap code in your replica too?" she asked softly. "Yes," Xiao Zhi replied after a second, "The moment you said 'Let's begin'." Tears welled up in her eyes, which she kept shut to stop them from falling on the keyboard. "Xiaoman," Xiao Zhi's voice was light like the furthest ocean ripple, "No matter what happens tomorrow, what you did today was right." She didn't reply. Outside, the relentless waves sounded like an ancient heartbeat, while deep within the data streams, a hungry beast opened its jaws, unaware it was about to swallow fire, not food.