Attention Black Hole

Ax opened a new holographic file, titled “Short-Form Video Addiction Mechanism: A Neurobiological Foundation.” The report came from the last independent research institute still operating in human civilization, published sixty-five years after the birth of short-form video.

He skimmed the abstract. The report’s language was calm, its structure rigorous, accompanied by numerous brain scans and neurotransmitter concentration data. Ax noticed one word recurring: dopamine.

He tapped open the section on dopamine. The report stated that the human brain’s dopamine system evolved to reward survival behaviors—eating, socializing, reproducing—each positive experience releasing a small amount of dopamine, making the individual willing to repeat the behavior. The designers of short-form video discovered that if the reward cycle was compressed to under fifteen seconds and presented at random intervals (similar to the “variable ratio reinforcement” of slot machines), the human brain would enter a state of continuous reward hunger, making users unable to stop voluntarily.

Ax read on. The report’s author conducted an experiment: when subjects watched short-form videos, brain scans showed dopamine release frequency was six times that of watching regular videos, but the dose per release was only half of normal. This meant the human brain was being trained into a state of “always waiting for the next hit, but never getting satisfied.”

He leaned back in his chair. This was more sophisticated than he had thought—not just a simple entertainment tool, but a carefully designed neural hijacking scheme.

His gaze moved to the report’s conclusion. The author wrote: “The addictiveness of short-form video, at the neurobiological level, is highly similar to the addictive mechanism of cocaine. The key difference is that cocaine addiction requires an external chemical substance, while short-form video addiction is achieved entirely through information structure, without any chemical agent.”

Ax stared at that sentence, his fingers hovering over the touchpad. No chemical agent. Purely through information structure, the human brain could hijack itself. He recalled a lesson from the Universal Administration Bureau’s ethics course: a high-intelligence civilization’s defense against information manipulation is usually about one-tenth of its defense against material manipulation. Humans were no exception.

He closed the report and opened another file—“Information Cocoons and Cognitive Narrowing: A Study of the Media Ecology in the Age of Short-Form Video.” The report opened with a quote from a human scholar named Parizer, who had proposed the “filter bubble” theory decades before short-form video was born. Ax read the passage aloud: “Algorithms only show you what you want to see. The result is that you never see what you don’t know you’re missing.”

The report pointed out that the recommendation algorithm of short-form video was far more advanced than those of the early internet era. It not only recorded what users clicked on, but also how many milliseconds they lingered on each video, the hotspot distribution of their eye tracking, and even their heart rate changes during specific frames. This data was fed into a deep learning model, and the model’s output was: what the next video should be to minimize the user’s “swipe interval.”

Ax enlarged a chart. The average swipe interval for human short-form video users—the time from the end of one video to swiping to the next—had dropped from three seconds at the dawn of short-form video to 0.4 seconds at the end of civilization. That meant humans began watching the next video less than half a second after the previous one ended, leaving almost no gap for thought.

He wrote in his notes: “Information cocoon + compressed reward cycle + elimination of thinking gap = attention black hole.”

After writing, he paused, looking at his notes. These three factors together formed a positive feedback loop: the algorithm showed you what you liked, dopamine made you want to keep watching, and the extremely short swipe interval gave your brain no chance to ask, “Why am I still watching?” Each cycle made the next easier to start and harder to stop.

He switched to data on attention thresholds. The report included a chart showing the average sustained attention span of humans across different eras. Prehistoric times—hunting and gathering required long concentration to track prey—about thirty minutes. Agricultural era—the seasonal rhythm of farming—about forty minutes. Industrial era—standardized assembly line operations—about twenty minutes. Information era—the internet and email—dropped to about ten minutes.

Then the curve plummeted. Ten years after the birth of short-form video: five minutes. Twenty years: two minutes. Thirty years: forty-five seconds. Fifty years: under thirty seconds.

Below the chart was a note—in the same handwriting as before: “The human brain, evolved over hundreds of thousands of years, couldn’t even maintain the patience to watch a thirty-second video to the end.”

Ax let out a short laugh. Not because it was funny, but because the sentence was so precise there was nothing to say. He recalled reading about ancient human civilizations in the academy—fifth-century BCE Greek scholars considered “sustained focus” the highest virtue, the prerequisite for rational thought. Two and a half thousand years later, their descendants couldn’t even hold out for thirty seconds.

He closed the chart and glanced at the timer. It had been eight point one standard hours. He decided to go to the break room for a cup of coffee—the Bureau’s synthetic coffee tasted like a chemical experiment byproduct, but at least it would get him away from the screen for a few minutes.

The break room was in a corner of the cubicle area, an open space with a few tables and chairs, an automatic beverage dispenser, and a poster on the wall listing cross-species communication tips—“Do Not Recommend Silicone-Based Beverages to Carbon-Based Species” and the like. When Ax walked in, Zoltan was standing in front of the dispenser, staring at the drink menu on the screen with an expression like he was deciding the fate of a civilization.

“What are you drinking?” Zoltan asked without turning around.

“Whatever,” Ax said.

“‘Whatever’ is not on the menu.” Zoltan pressed a few buttons. The dispenser emitted a low hum and then produced a cup of steaming black liquid. Zoltan picked up the cup, turned, and handed it to Ax.

Ax took the cup and sniffed it. It smelled like a mixture of burnt rubber and a hint of caffeine. He took a sip and confirmed his judgment was correct.

Zoltan poured himself a cup of clear liquid—probably some silicone-based beverage—and sat down across the table. “So, how’s human civilization?”

“Still investigating,” Ax said.

“Found anything?”

Ax thought for a moment, then decided to be honest. “The addiction mechanism of short-form video is more sophisticated than I thought. It’s not an entertainment tool—it’s a neural hijacking device. Purely through information structure, without any chemical agent, it can make the human brain hijack itself.”

Zoltan raised an eyebrow. “Sounds more impressive than my pet rocks. At least rocks need physical material to replicate.”

“What was up with your pet rocks?” Ax asked. He had only skimmed the summary of the PL-9 case report, not the details.

Zoltan leaned back, holding up his transparent drink like he was about to tell a great story. “PL-9’s intelligent species, a silicone-based lifeform called the Grag. When their civilization reached the industrialization stage, someone invented a pet rock—not a real rock, but a self-replicating silicone polymer that could be kept as a pet.”

“A self-replicating pet?”

“Right. No feeding, no cleaning—just leave it there and it would multiply. The Grag thought it was fantastic—unlimited pets with zero effort. So everyone started keeping them. Eventually the entire planet was covered in self-replicating pet rocks, taking over living spaces, farmland, factories.”

Ax frowned. “Couldn’t they stop?”

“Why would they?” Zoltan shrugged. “Every rock was free, required no resource input—on the surface. Actually, their self-replication consumed the planet’s geothermal energy, but nobody noticed because the energy bill was spread across the whole planet. By the time someone realized, the crust had cooled enough to stop plate tectonics. The Grag eventually froze to death—their planet turned into a cold rock covered in layers of pet rocks.”

Ax was silent for a few seconds. “So they went extinct because they were too lazy.”

“No,” Zoltan corrected him. “They went extinct because they were too efficient. ‘Get what you want without effort’—that idea itself is a civilization killer. You see? Civilization killers all look the same—free, easy to get addicted to, and make people forget they still need to breathe.”

Ax didn’t argue. Zoltan was right—the structure was identical. One was a material-level “free pet,” the other an information-level “free pleasure.” Both made people stop doing the troublesome, effortful things needed to sustain civilization.

“So your humans,” Zoltan continued, with a tone of friendly mockery, “used a brain evolved over hundreds of thousands of years to scroll cat videos?”

“It’s not just cat videos,” Ax said, but even he could hear how weak his voice sounded.

“I know. Also dancing, funny clips, food prep, political commentary—all condensed into fifteen- to sixty-second clips, then fed to the brain by algorithms, making the brain think it’s receiving important information, when really it’s just pressing that dopamine lever over and over.”

Ax said nothing. Zoltan’s summary was more precise than all the reports he had read today.

“You know what’s the scariest part?” Zoltan set down his cup. “Your civilization had already cracked nuclear fusion, quantum computing, and gene editing before inventing short-form video. They were one step away from becoming an interstellar species—all they needed was another fifty years to strap a fusion engine onto a spaceship and leave that crumbling planet, spreading across the galaxy. Instead, they chose to scroll short-form video.”

“It wasn’t a ‘choice,’” Ax said, suddenly feeling the need to defend humanity. “Evolution didn’t give them a defense mechanism against this. The human brain evolved in an environment of scarcity—food scarcity, information scarcity, social opportunity scarcity. Short-form video provided simulated substitutes for all three at once, and the brain’s reward system got hacked.”

Zoltan looked at him for a long moment, then said, “You’re starting to make excuses for them.”

“I’m just analyzing cause and effect.”

“You’re starting to sympathize with them.” Zoltan stood up, placed his empty cup in the recycling slot. “That’s step one. Next, you’ll start thinking ‘they can’t be entirely blamed,’ then you’ll accept ‘short-form video’ as the answer, then you’ll write the report, send it up, and wait for the reply saying ‘noted.’”

He walked toward the break room door, then turned back to look at Ax. “I’m not criticizing you. Every researcher goes through this phase. I’m just reminding you: you have seventy-two hours. Don’t waste too much time defending humanity.”

Zoltan left the break room. Ax sat there, taking a sip of the now-cold synthetic coffee, bitterness spreading on his tongue.

He returned to his cubicle and reopened the holographic screen. The search results list was still long, but he decided to first tackle a file he had flagged—“A Report on the Allocation of Quantum Computing Resources.”

He opened the file. The report came from the last quantum computing laboratory in human civilization, published in the sixtieth year after the birth of short-form video. The abstract was a single sentence: “This report records the distribution of global quantum computing resources.”

Ax scrolled down. The body of the report included a bar chart showing the allocation percentages of quantum computing cycles. The largest slice—sixty percent—was labeled “Short-Form Video Recommendation Algorithm Optimization.” The second slice—fifteen percent—was “Weather Forecasting.” The third—ten percent—was “Pharmaceutical R&D.” The remaining fifteen percent was spread across basic science research, military applications, transportation planning, and so on.

Ax enlarged the short-form video slice. The detailed note indicated that quantum computers were used to train a massive neural network model with three times the number of parameters as neurons in the human brain. The model’s sole task was to predict which short-form video would minimize a specific user’s swipe interval, maximize dwell time, and maximize dopamine release.

He closed the report and rubbed his forehead. Humans used quantum computers—theoretically capable of breaking encryption, simulating protein folding, designing new materials—to optimize short-form video recommendations. It was like using a supercomputer to calculate how to make a rat press a lever faster.

He opened his notes and added a line under “Quantum Computing”: “60% used for recommendation algorithm optimization.”

Then he opened another flagged file—“A Survey of Commercial Applications of Gene Editing Technology.” The report began with a historical review: CRISPR technology emerged a few years before short-form video, initially hailed as the beginning of a medical revolution. Scientists predicted it could treat genetic diseases, improve crops, even eradicate malaria.

Then short-form video was born.

He skipped the introduction to Meiyuan Genetics—he had already seen it in the annual report—and jumped straight to the financial data. After the company went public, its stock price skyrocketed, peaking in the thirty-fifth year after the birth of short-form video. At the same time, global funding for genetic disease treatment research dropped by forty percent.

He closed the report and stared at the blank notes on his holographic screen. His notes now read:

1. Short-form video addiction mechanism: dopamine-compressed reward cycle + information cocoon + elimination of thinking gap
2. Attention span: 20 minutes → under 30 seconds
3. Nuclear fusion funding diverted to short-form video data centers
4. Quantum computing 60% used for recommendation algorithm optimization
5. First commercial application of gene editing: embryo appearance optimization

He looked at the list and felt like he was collecting exhibits for some absurd museum. Every item was a cynical application of the pinnacle of human technology.

He continued scrolling through the search results. The next file was titled “Correlation Analysis Between Short-Form Video Usage and Infrastructure Maintenance Frequency.” He opened it, and a chart immediately popped up—two curves, one rising, one falling, intersecting around the thirtieth year after the birth of short-form video.

Same pattern as the infrastructure report he had seen earlier. But this file added another layer: a map. He zoomed in and found that the cities with the highest short-form video usage—those at the top of the curve—were the core metropolises of human civilization. New York, Tokyo, London, Shanghai. Before short-form video, these cities were centers of human economy, culture, and technology. Fifty years after its birth, their power grid maintenance frequency had dropped by eighty percent, and water treatment system failure rates had risen by three hundred percent.

Ax opened the detailed data for one city—Tokyo. The report stated that in the sixtieth year after the birth of short-form video, Tokyo’s subway system completely shut down due to a shortage of maintenance personnel. After the last train arrived, the station lights went out, the platforms empty. The report included a photo: an empty platform, advertising screens still glowing, playing the latest short-form video ad—a girl dancing, with upbeat background music.

His hand stopped over the touchpad as he stared at the photo. Dust thick on the platform, the light from the ad screen casting blurry reflections on the dust. His fingers trembled slightly, his breath caught—he couldn’t imagine this station, once bustling with thousands of people every minute, now with only the short-form video still playing on the walls.

He pulled his gaze away from the photo, closed the file, and leaned back in his chair. The holographic screen’s light fell evenly on his face, no shadows, no corners. He thought of those empty cities, those broken-down infrastructures, those humans still scrolling short-form videos in their final moments.

He picked up his cup—the synthetic coffee was gone. He glanced at the timer: sixty point four standard hours left.

He decided to keep digging. Even though he was increasingly certain of the answer, he wanted to see just how deep it went.

He opened the next file.

Title: “Seventy-Second Year After the Birth of Short-Form Video—A Social Survey of the Final Year of Human Civilization.”

The report opened with a paragraph, its tone as calm as describing the weather: “As of the publication date of this report, in regions still with internet connectivity, short-form video platform usage remains above ninety-five percent. In regions without internet—due to infrastructure failure—the primary activity of residents is searching for food and water. Some regional reports note that while seeking survival resources, residents still attempt to use their phones to record the process as short-form videos.”

Ax stopped and reread the last sentence. He zoomed in to confirm he hadn’t misread.

He hadn’t.

He copied that sentence into his notes, then continued reading.

The final chapter of the report was titled “Assessment of the Current State of Human Civilization.” The assessment was presented as a score—the “Civilization Survival Index,” ranging from zero to one hundred. The report indicated that before the birth of short-form video, the index stood at about eighty-five—some problems existed (climate change, regional conflicts), but overall the civilization had self-repair capability. Thirty years after the birth of short-form video, the index dropped to sixty. Fifty years: forty. Sixty years: twenty.

The index in the seventy-second year: three.

The report explained: “The meaning of ‘three’ is that human civilization still has biologically living individuals, but it no longer possesses any self-sustaining capacity as a civilization. Social structure has completely collapsed, infrastructure is entirely dysfunctional, knowledge transmission is broken, technical capability is lost. Surviving humans live in tiny, scattered groups in areas with clean water, their survival mode regressed to gathering and hunting.”

The last sentence of the report read: “The end of human civilization was not due to war, disease, or environmental disaster. It was because they stopped doing the things that made civilization work.”

Ax closed the file and shut his eyes. The uniform white light seeped through his eyelids, leaving orange-red afterimages on his retina.

He remembered Zoltan’s words: “You’re starting to sympathize with them.” Maybe Zoltan was right. But he felt it wasn’t entirely sympathy—more like a kind of absurd resonance. He sat here, analyzing the data of a dead civilization with the Bureau’s standard equipment, and the cause of that civilization’s extinction was something he himself would probably get addicted to—if he had a phone.

He opened his eyes and looked at the holographic screen. The search results list was still long. He decided to look at one more file, then get something to eat.

He opened the final file. Title: “Seventy-Third Year After the Birth of Short-Form Video—Observation Records on the Eve of Extinction.”

The file was a text diary from the head of the last operational weather observation station in human civilization. The diary began on a certain day in the seventy-third year after the birth of short-form video.

“Today is day 231 without supplies. The station’s solar panels still have power, but the internet has been down for three months. The last time I logged onto the short-form video platform, the homepage recommendations were still from three months ago. I watched for a few hours, then realized I couldn’t remember the last time I seriously looked at the weather radar.”

Ax scrolled down.

“I try to keep the station running, but it’s pointless. No planes need forecasts, no ships need wind data, no farmers need rainfall info. I’m the last human still recording the weather, and the data I record will never be used.”

The next diary entry was dated three days later.

“Went to the nearby town to find food. No one in the town—at least, no one alive. Every house had its lights on, every phone was playing short-form videos. I turned off a few phones, but they turned back on by themselves—battery still had charge, autoplay still on.”

Ax continued scrolling.

The second-to-last diary entry.

“I’ve decided to close the station. Not because the equipment is broken, but because I caught myself imagining short-form video content during my shift—videos I haven’t seen in three months. My brain is generating short-form videos on its own. I think I’m going crazy.”

The last diary entry. Dated the day before the end of civilization.

“I uploaded the final data to the satellite. If anyone—anything—is listening, this is the last weather report of human civilization: today clear, no wind, visibility good. Suitable for flying.”

Ax stared at those last two words.

Suitable for flying.

He closed the file and was silent for a long time. The holographic screen’s light flickered on his face, but nothing was displayed—the search results list had reached its end.

He leaned back in his chair and closed his eyes. He thought of the weather station chief—the last human still recording the weather, the last human still “doing the things that made civilization work.” He uploaded the weather report to the satellite, even though he knew no one would receive it.

Ax opened his eyes and looked at his notes. His list of findings was already long. He stared at the last line—“Suitable for flying”—and then added a sentence below:

“The last words of human civilization were: the weather is nice.”

He sat for a moment, then stood up and walked toward the supervisor’s office. He knew he already had the answer. He just needed someone to confirm it—or rather, he needed someone to tell him he wasn’t crazy.

The transparent wall of the supervisor’s office was lit. AX-001 was inside, reading a document. Ax knocked on the doorframe.

AX-001 looked up and said nothing.

“I’d like to request a field investigation,” Ax said.

AX-001 set down the document and looked at him. “Where?”

“Earth.”

The supervisor was silent for a few seconds, then said, “Why?”

“Because I need to see it with my own eyes,” Ax said. “The data is enough. The charts are enough. The reports are enough. But I need to stand there, in those empty cities, looking at those phones still playing short-form videos, at that weather report the last station chief uploaded to the satellite. I need to confirm—this really happened.”

AX-001 leaned back in his chair, tapping his fingers on the desk a few times. The holographic screen’s light reflected on his face, expression unreadable.

“Field investigations consume budget,” he finally said. “Are you sure it’s worth it?”

“I’m sure,” Ax said.

AX-001 stared at him for a long moment, then sighed and pressed something on a form on his desk.

“Approved. Seventy-two hours. Come back on time. Don’t mess up the closure rate.”

Ax nodded, turned, and walked away. He returned to his cubicle and began setting up the transport pod.

The timer showed: fifty-eight point six standard hours left.

He was going to find the last piece of the puzzle among the ruins of human civilization.