I'm doing engineering in the instance.

Xie Chengzhou looked at the water level figures a second time.

5.46 meters. 2.46 meters from the warning line.

They spent two fifteen-minute intervals in the corridor. The rhythmic sounds of the seepage outside stopped, but stopping didn't mean disappearing; they just couldn't hear them anymore.

The maintenance room is not the end.

He shone his flashlight on the metal panel, which was about 60 centimeters wide and 1.2 meters high. It had four new, silver-white bolts without rust, a completely different color from the old bolts on the surrounding walls. Someone had recently replaced them.

He said, "Let's go into the expansion joint passage."

Engineer Li had no objections. His left ring and little fingers were still recovering, but he switched the flashlight to his right hand, adjusted his posture, and stood next to Xie Chengzhou.

Old Zhao took a sip of hot water and put the lid on his thermos. "How wide is the passageway?"

"60 centimeters," Xie Chengzhou said, "enough for one person to pass sideways."

Old Zhao nodded without saying anything more.

Xie Chengzhou took out his tools and began to remove the bolts. The bolts were newly replaced, and they were easier to remove than he had expected. He removed all four bolts within three minutes. He pulled the metal cover plate outward. The cover plate was heavier than he had expected. It was made of steel, not aluminum alloy. He had to use both hands to move it and lean it against the wall.

There is airflow in the passage.

It's not big, but it exists.

He held his flashlight inside and shone it around—the passageway was made of concrete, with smooth walls on both sides, water seepage marks on the ceiling, and a dry floor without any standing water. The passageway stretched on and on, but the flashlight beam couldn't reach the end.

He noticed something.

There were footprints on the ground of the passageway.

These aren't water stains; they're genuine shoe sole marks, pressed into the dust, and pointing in both directions—they've been in and out. The sole tread is clear, indicating they're construction site safety shoes, not athletic shoes.

He wrote a line in his memo: "Expansion joint channel · P2-07 · Footprint · Two-way · Safety shoe sole pattern · Source: Construction worker or player · Time: Unknown · Intent: Unknown."

Then he went in sideways.

The passage was longer than he had imagined.

He counted his steps, each about 60 centimeters long. When he reached the eighteenth step, the passage turned to the left at an angle of about 30 degrees. It wasn't a right angle, but a gentle curve. This wasn't the normal direction of an expansion joint—expansion joints should be straight, following the direction of building expansion and contraction, and shouldn't have gentle curves.

He stopped.

The flashlight shone on the wall inside the bend.

There is a mark on the concrete surface, about 1.5 meters high, extending horizontally, as if something scraped against it. It is about 3 centimeters wide and very shallow; it is a scratch, not a crack.

He went through it in his mind: the scratch was 1.5 meters high, 3 centimeters wide, and horizontal in direction. The source was probably the friction between the object and the wall when moving it.

Someone has been moving things in this passageway.

He suppressed this judgment and continued walking inside.

In the twenty-seventh step, he smelled the scent.

It wasn't a musty smell, nor a fishy smell; it was a familiar odor—quicklime.

He had smelled this odor on the construction site before; it was used for foundation treatment during the basic construction phase, and also during the construction of the building's damp-proof layer, and also—

It is also used for crack filling treatment.

Someone has filled in this passageway.

He shone his flashlight forward, and when he reached about 30 steps, he saw a small area on the ground of the passageway. The area was lighter in color than the surrounding area and was filled with new filling material. It was about 0.3 square meters in size and had irregular edges. It was applied by hand, not by machine. The application was uneven and there were traces of overflow at the edges.

Xie Chengzhou squatted down and touched the edge of the filled area with his finger.

It's hard. It's already hardened; it's not newly constructed, but rather has been there for some time, though not too long. There are no signs of the surface getting damp again.

他在备忘录里写：「通道内·地面填充区域·P2-07-A·面积约0.3m²·材料：水泥砂浆+石灰·固化状态：完成·时间推断：7-30天内·施工方式：人工·施工质量：不均匀·目的：未知。」

He stood up and continued walking inside.

Engineer Li followed behind him, while Old Zhao brought up the rear.

The sound of three people's footsteps echoed clearly in the concrete passageway. Xie Chengzhou listened for a while and then softened his own steps.

He stopped at the thirty-fifth step.

There is a bulge in the ground at the thirty-fifth step of the passage.

It's small, about 5 centimeters high and 20 centimeters wide, spanning the entire width of the passageway. It looks like a localized bulge caused by uneven stress on the ground. The concrete surface has tiny radial cracks, extending from the center of the bulge outwards. The cracks are about 0.5 millimeters wide, and the depth is not visible.

His "structural perspective" provided the conclusion at this moment: foundation settlement.

The judgment was quick, almost instinctive—the radial distribution of the cracks, the location of the bulge, and the overall stress state of the channel all pointed to the same conclusion: localized foundation settlement, surface layer cracking under pressure, and the bulge being the result of the rebound of the underlying soil after water loss and shrinkage.

His right foot was already raised.

He has to cross over.

Then he stopped.

It wasn't because someone told him to stop. It was because he stopped on his own.

He didn't know why he stopped; something just got stuck in his mind, like a thin thread tightening, but he couldn't find the source of that thread.

He put his right foot back, squatted down, and brought the flashlight closer to examine the bulge up close.

The radial distribution of the cracks is correct.

The bulge is in the correct location.

But there's one detail that's wrong.

The edges of the crack were too clean.

Cracks caused by foundation settlement should have powder flakes and tiny debris at the edges, and should have the rough texture of concrete naturally breaking under stress. But the edges of this crack are smooth, as if it was cut with a tool, rather than cracked under stress.

He adjusted the angle of the flashlight and shone it from the side.

There is a slit at the bottom of the bulge.

It's not a crack, it's a fissure, about 1 millimeter wide, uniform, extending along the raised edge, like—

It looks like a lid.

His hand did not move.

He mentally reviewed the structure: the ground was raised, with evenly spaced gaps along the edges, and the surface cracks were artificially cut, not naturally formed. This wasn't ground subsidence; it was a man-made trigger mechanism. The cracks were a camouflage, the raised area was a cover, and what lay beneath the cover—he didn't know, but he knew what would happen if he had stepped over it and stepped on that cover.

He didn't know what would happen, but he knew it was a trap.

In his memo, he wrote: "Trigger device P2-07-B Location: Step 35 of the passage Camouflage: Foundation settlement Identification: Smooth crack edges Judgment: Man-made Trigger method: Stepping on Result: Unknown Identified, not triggered."

Then he paused for a long time after that line of text.

He was thinking about something.

His "structural perspective" led to the conclusion of "foundation settlement," but that conclusion was wrong.

This is the second time.

The first one was man-made, and his intuition told him to give it high priority, but Lao Zhao told him to stop.

This time it was the trigger device. His intuition told him that the foundation was settling, and he stopped at the last moment, but he didn't know why he stopped—he didn't see the problem at the edge of the crack. He stopped first and then saw it.

Stop first, then look.

It wasn't intuition that gave him the judgment, but rather intuition that caused him to pause.

He went through the difference in his mind, but couldn't quite figure it out. However, he felt that the difference was very important.

He added a line to his memo: "Second deviation in 'structural perspective': The triggering device was identified as foundation settlement. Source of the pause: Unknown (not an intuitive judgment, but an intuitive pause). Mechanism: To be investigated."

He stood up and said to Li, the engineer behind him, "There's something here. Walk around it, stay close to the right side wall, and don't step in the middle."

The passage is only 60 centimeters wide. "Right side wall" means to stick to the right wall and move around the area by sidestepping. With your right shoulder against the wall, step in with your left foot first, then your right foot, without stepping on the center line of the passage.

After listening, Engineer Li turned to the side and walked over as instructed.

Old Zhao walked over there, glanced down at the bulge, didn't ask any questions, and walked around it.

Xie Chengzhou continued walking forward.

On the fortieth step, he heard the sound of water.

It wasn't the sound of seeping water, but the sound of flowing water, with velocity and pressure, coming from the wall on the right side of the passage. The sound was steady and continuous, not intermittent.

He shone his flashlight on the right-hand wall.

There is a damp stain on the wall surface, extending from the top to the bottom, about 15 centimeters wide. There are white calcium carbonate deposits at the edges of the damp stain, which are left by long-term water seepage. It is not today, but has been seeping for a long time.

But the sound of water is new.

He had distinguished this sound on the construction site—the old seepage point suddenly increased in pressure, not naturally, but due to the increased pressure difference caused by the rise in the external water level.

He glanced at his wrist.

5.51 m.

In five minutes, the water level rose by 0.05 meters.

He did some mental calculations: the speed was faster than before. Before, it was about 1.44 meters per hour, but now it was 0.01 meters per minute, which translates to 0.6 meters per hour.

No, it's not that it accelerated; the problem lies in his calculation benchmark. The previous figure was a 250-minute safety window, which was an estimate at the time of entry, not real-time data. The actual rate of water level rise is constantly changing, and he doesn't have enough data points to judge the trend.

He wrote this uncertainty into his memo and then moved on.

Step 50: The passage has widened.

It didn't widen suddenly, but slowly, expanding from 60 centimeters to about 1.2 meters, and then to 1.8 meters. Xie Chengzhou realized that they had walked out of the expansion joint passage and entered a new space.

He swept the flashlight around.

It is an equipment room, smaller than the maintenance room of P2-07, about 15 square meters. There are pipes, valves, pressure gauges on the walls, and a distribution box. The door of the distribution box is open, and the wiring inside has been moved. Some of the wiring has been cut, but the cuts are neat and were made with tools, not from aging or breakage.

Someone damaged the wiring between these devices.

Xie Chengzhou stood there, putting this equipment room together with what he had seen before—the man-made cracks, the filling work in the passageway, the triggering device, and the cut wires.

This wasn't done by one person.

This is a complete sabotage plan, with a specific plan and steps, and someone has made arrangements in advance in this instance.

In his memo, he wrote: "Systemic sabotage • P2-07 area • Evidence: ① Man-made cracks • ② Channel filling construction • ③ Triggering device • ④ Line damage • Judgment: Organized premeditated behavior • Purpose: Unknown • Perpetrator: Unknown • This is not the natural rule of copying; this is someone's interference with copying."

Then he heard a sound.

A voice came from the direction they came in, from the expansion joint passage.

It's not a rhythmic sound.

It's the sound of dragging.

It felt like something was sliding on a concrete surface; it was heavy, there was resistance, but it kept moving.

His flashlight was immediately turned towards the entrance of the passage.

You can't see anything in the passage. When you shine a flashlight in, it only illuminates about ten steps. After the tenth step, there's a bend, and you can't see anything after that.

But the sound was getting closer.

He said, "Where is the exit?"

It wasn't a question, it was a statement; he was scanning the four walls of the equipment room, looking for an exit.

There was a door in the equipment room, opposite the direction they came from. It was an iron door, rusted, and there was light coming from the crack in the door—not the light of a flashlight, but a different kind of light, colder and more even, as if it were coming from outside.

Li Gong had already walked to the door, placed his hand on the doorknob, and glanced back at Xie Chengzhou.

Xie Chengzhou said, "Open."

The iron gate made a dull thud, its rust jamming it. Li Gong nudged it with his shoulder, and the gate opened, letting in a rush of cold air. Xie Chengzhou smelled water, a lot of water, the smell of an open body of water, not the smell of pipes.

They went outside.

The outside is the top of the dam.

The concrete surface of the dam crest is about 8 meters wide, with guardrails on both sides. To the left is the reservoir, and to the right is the downstream side of the dam. The water level downstream is much lower than upstream, with a height difference estimated to be over 20 meters. The sky is gray, there is no sun, and the wind is strong, blowing from upstream and carrying the smell of water.

Xie Chengzhou glanced back at the iron gate.

The door closed by itself after they came out; the wind pressure held it shut.

The dragging sound in the passageway was still there, but through the iron gate, it became a low background noise, like the sound of machinery running in the distance. If you didn't know what it was, it would be easy to ignore.

Old Zhao stood next to Xie Chengzhou, unscrewed the lid of his thermos, took a sip, screwed it back on, and put it back in his hand.

"It's out," he said.

Xie Chengzhou did not answer immediately.

He was looking at the ground on top of the dam.

There are cracks on the dam crest surface, not just one, but many, with irregular directions and widths ranging from 0.5 mm to 3 mm. Some cracks have water seeping out, while others are dry. This is normal aging of the dam crest surface layer, not a structural problem, and does not require treatment.

His "structural perspective," while scanning these cracks, provided a series of judgments: this is a temperature crack, this is a shrinkage crack, this is—

He stopped.

He was thinking about something.

His "structural perspective" scans the cracks, determines their causes, and assesses their hazard level—all very quickly and automatically, like breathing, requiring no conscious activation.

But what is the underlying logic of this ability?

Its judgment is based on the fact that the shape of the crack corresponds to the result of natural forces. Temperature changes, shrinkage stress, exceeding the bearing limit, and foundation settlement—these are its reference points, and all of these are natural forces.

It lacks a frame of reference called "deliberate human intervention".

He had seen too many cracks on construction sites. In his experience, cracks originated from natural forces, and occasionally from construction quality issues, but those issues were unintentional, mistakes, not deliberate. His intuition had never been trained to consider the possibility that "someone deliberately created a crack here."

So when the man-made crack appeared, his intuition led him to judge it as a "high-priority node," which was the correct judgment—the crack was indeed of high priority, but the reason was wrong; he thought it was a structural problem, not a man-made trap.

When the trigger appeared, his intuition gave him the judgment of "foundation settlement," which was wrong. He stopped not because his intuition told him there was danger, but because his intuition gave him a judgment, but something did not match the judgment with reality. He felt that "discrepancy," but he did not know what it was. He just stopped.

His intuition could sense that "something was wrong," but it couldn't tell him that "it was done by a human."

This is a blind spot.

It's not that intuition has failed, but that the variable of "human" is missing from the frame of reference of intuition.

He wrote in his memo:

"Structural Perspective" Blind Spot Confirmation: The reference frame is natural forces plus construction errors, excluding the variable of "deliberate human sabotage". In scenarios with premeditated intervention, the judgment results will be systematically biased—not randomly biased, but directional biased: mistaking man-made traps for natural structural problems, and mistaking artificial triggering devices for foundation settlement.

This blind spot cannot be compensated for by "observing more carefully," because the observation itself is conducted within a flawed frame of reference.

Solution: Before proceeding with the judgment, ask yourself this question: Has anyone been here? If so, what did they do?

It's not that engineering problems come first, it's that human problems come first.

He closed the memo.

The wind was blowing from upstream, and he felt the numbness in his right ankle.

It's still there, but a little weaker than before.

Li Gong stood next to him, bent the ring finger of his left hand halfway, stopped, and then stretched it out.

"I'm feeling a little better," he said. "I can feel it now, but it's still too late."

Xie Chengzhou nodded.

Old Zhao was looking at the railing on the top of the dam. He tapped on it, listened to the sound, and said, "The railing is hollow and rusted through. You can't lean against it."

Xie Chengzhou glanced at the numbers on his wrist.

5.58 m.

Distance from the warning line: 2.42 meters.

They spent about ten minutes in the passageway.

The wind was strong at the top of the dam, blowing his work clothes tightly against his body. He mentally reviewed the current situation: three people, dam top, water level 5.58 meters, 2.42 meters from the warning line, seepage person in the passage, man-made cracks not treated, triggering device not activated, expansion joint passage entrance has been left, and the impact of line damage is unknown.

There was still one thing he hadn't figured out.

Who did the destruction? What was their purpose? Are they still in this instance?

He wrote the following line in his memo: "Unresolved issues: Spoiler identity, purpose, and current location. If they're still in the copy, are we now part of their plan, or outside of it?"

He closed the memo, looked up, and gazed at the direction the dam crest extended outwards.

The dam crest is very long, disappearing into the gray light at both ends, with no end in sight.

They need to move forward.