POST AGI
Sep 26, 2026
1. Utopia
The morning light of 2176 looked no different than it did a century ago. I sat in the climate-controlled dining room eating breakfast, the edges of my retinal projection quietly scrolling through today’s demographic bulletin: a global population of 6.5 billion, declining for the 131st consecutive year.
In this era, this was just background noise, as mundane as a weather forecast. In the next room, my mother was calibrating her deep-dive equipment. She planned to travel beneath the ice sheets of Europa next month and was cheerfully using the voice interface to pick out the color of her pressure suit. No one felt any tension over the fact that I was attending my "twenty-five-year final review" this afternoon; my friend was even in my comms channel, excitedly discussing the underground performance we were going to see tonight.
People of the past harbored an arrogant illusion. They believed that once food, housing, and disease were completely eradicated, humanity would multiply like bacteria in a petri dish. Later, we really did solve all of those things—and the population began an irreversible decline.
In this world, extreme abundance makes raising a child incredibly easy, but it makes not having one equally effortless. There is no anxiety over survival, no fear of old age, and certainly no pressure to carry on any sort of social capital. People can delay childbirth until well past their hundredth birthday, or, like the generations before my mother, they can simply never choose it at all. As lifespans stretched, time became something rich and weightless.
This society is, indeed, worth living in.
2. Humanity
Before my final review that afternoon, I went with my friend—who had already opted out of the system—to watch an unrestricted free-climbing competition.
The venue sat beneath a natural rock face hundreds of meters high. There were no safety ropes; just human flesh pitted against gravity. In truth, if bionic robots equipped with kinematic algorithms were sent to climb it, they could map out an absolutely flawless route, their movements precise to the micrometer. Yet people were still willing to pay a premium to watch a real human.
Because the girl on the cliff would actually get tired. Her muscles would involuntarily tremble in the wind, real terror would flash in her pupils, and she truly could slip and fall on her very next reach. The audience held their breath because everyone knew she couldn't just swap out her body. Perfection is cheap today; fragility and failure are the rare luxuries.
My friend watched the girl on the precipice, took a sip of his synthetic beverage, his eyes bright. He had exited the cultivation system two years ago and now lived an incredibly fulfilling life: studying classical art, delivering public lectures everywhere, and even building his own micro-community. He never made "proving the AI's cognitive limits" his life's mission.
Society today is like a perfectly, softly sliced cake. The vast majority are free citizens like him, enjoying endless life and liberty. A tiny minority are athletes and celebrities who retain market demand because they provide authentic emotional value. And I belong to the most marginal sliver—those still in the cultivation system, trying to find something.
I respect his choice; it’s a transparent, lucid kind of happiness. But deep down, I still feel a roaring train waiting for me in the distance. I am running toward something heavier, something more important.
3. The Great Filter
The cultivation system has a better-known name in this era, one tinged with classical tragedy: The Great Filter.
The filtering starts very early. My cohort originally had 13 million people under observation. At age ten, the system conducted its first gentle guidance, leaving 650,000. At age fifteen, as academic depth and mental models grew infinitely more complex, only 130,000 remained.
Every departure is not an elimination, but a "withdrawal." Throughout this agonizingly long process, anyone can press the withdrawal button at any time. The system grants a one-year grace period. If you don't retract it within a year, you graduate early as a free citizen to enjoy that flawless utopia outside. Some leave early; others simply find the high-intensity knowledge absorption fun and are willing to treat it as a game until they turn twenty-five.
I made it all the way to today. From childhood, the AI has been our mentor, nanny, and mirror. It taught us the underlying logic of the world and helped us chart our own talents and limits with terrifying precision. Because of this, the vast majority of people trust its judgment implicitly. For centuries, not a single person deemed "advised to withdraw" has ever left and gone on to make a disruptive achievement that overturned the AI's verdict. This absolute accuracy makes the filter system as unshakeable as bedrock.
At dusk, the results of the final review popped up on my retina.
Between voluntary withdrawals and today’s final computational assessment, 99.99% of the remaining candidates had been stripped away. In this cohort, exactly thirteen people received the invitation to continue cultivation.
I was one of them. And at the edge of the list, I saw a familiar name: Young. A peer I occasionally bumped into in the cross-disciplinary seminar rooms over the past few years, someone who always just sat silently listening to others speak.
4. One Hundred Gigawatts
This was my first time, as an official candidate, being granted permission to access the system’s foundational, massive computing and experimental facilities.
Free citizens can, of course, study all known knowledge, debate esoteric philosophy or physics with the AI all night, and even publish their own theories in virtual sandboxes. But they cannot, simply out of "personal curiosity," apply to use a deep-sea particle accelerator or invoke planetary-scale computing nodes.
I didn't receive better material comforts or any so-called privileges for becoming one of the thirteen. In this system, only when you actually produce a breakthrough will your contribution be recognized based on its real economic and social value.
But the moment my clearance was approved—when I fed a half-formed, complex topological conjecture into the neural interface and watched the panel display: “100-Gigawatt facility allocated, verification in progress”—I still felt an overwhelming, electric shudder that was impossible to resist.
At this very moment, a fleeting thought of mine was causing the cooling waters of some ocean thousands of miles away to violently heat up.
5. Independent Discovery
Five months later, I achieved my first result that made my palms sweat.
It wasn't some grand unified theory set to upend relativity or reconstruct cosmology. Its entry point was tiny: when calculating the collapse probability of a multi-dimensional quantum entangled state, traditional algorithms always hit an exponential computational dead end. I seemed to have found an exquisitely subtle path, bypassing the obstacle through the crevices of a high-dimensional manifold.
I verified it fourteen times. Every single derivation formed a perfect closed loop. The result held.
Taking a deep breath, I submitted the complete mathematical proof to the AI.
In less than a second, the AI's gentle voice echoed in my mind, confirming my derivation was flawless. But immediately after, it unfolded a document in my field of vision.
It was a secondary model constructed by a previous generation of researchers over eighty years ago. My method was mathematically entirely equivalent to the core concept of that paper; I had just used a completely different, even slightly clumsier, geometric expression.
The AI had no intention of mocking me. Its voice remained calm and affirming: "The fact that you could rely on your own intuition to cross this mountain and arrive here independently proves that you have truly comprehended the foundational structure of this problem."
I looked at that eighty-year-old paper, my heartbeat slowly returning to normal. I accepted the explanation. The first failure did, in fact, feel very much like a part of success. At the very least, I had proven I had the ability to touch the wall of humanity's known boundaries.
6. Another Path
I began to pivot, actively steering clear of mature fields that had been repeatedly steamrolled by predecessors.
I spent a massive amount of time reading fringe materials across different disciplines, searching for variables that had never been discussed in the same room. During this process, Young and I crossed paths in the virtual hub’s physics sandbox more and more often. Gradually, Young became my regular sounding board.
We would take walks, debating the boundaries of information entropy in a simulated Cretaceous jungle, or using our toes to trace complex formulas in the silent dust of the Martian surface. Young didn't speak much, but every word was razor-sharp. Sometimes, a computational bottleneck I'd been stuck on for weeks would instantly shatter just because of a casual, offhand remark from Young.
Early one morning a year later, I once again thought I had caught something new. It was a cross-solution regarding dark matter density fluctuations and non-standard gravitational wave models.
I submitted the result with uncontrollable excitement.
This time, the feedback the AI returned was a set of code and mathematical language I couldn't read at all. It took me three full months, with the AI's assistance, to slowly chew through the foundational logic of that language. When the fog finally cleared, I stared at the final equation, feeling like I had plunged into an ice bath.
The theory not only fully contained my discovery, but it was infinitely deeper and broader. My finding was merely a special case of that colossal model, locked in an extremely low dimension.
My frustration pierced me deeply: in my own field of research, I had to learn an entirely new language just to barely understand why I had actually discovered absolutely nothing.
7. Year Seventy-One
I halted all my experiments and began spending my nights pouring over the research logs of past Homo sapiens, particularly the year 2105.
That was the last time in human history that a theoretical breakthrough was driven purely by human intellect. Since then, for exactly 71 years, not a single human being had been able to leave "something new" in this world.
To the free citizens outside, "71 years without a new human discovery" was a painless piece of historical trivia. It didn't affect their lives; the AI smoothly drove technological iteration, and society ran perfectly. But to those of us left in the institute, that number was a stopwatch hanging over our heads, one that could never be reset to zero. It ticked endlessly, mocking our efforts.
I wanted to find some replicable methodology from the experiences of our predecessors. Through this anxious process, my relationship with Young deepened. We were like two fish diving alone in the abyss, the only ones who could hear each other’s sonar.
Young's gaze always seemed fixed on something further away, something more void than what I was looking at. Young's research began to undergo highly abstract shifts, and even I struggled to keep up with that kind of non-linear thinking. But neither of us called out the change. I stubbornly believed we were just two comrades standing back-to-back in this massive labyrinth, each looking for our own exit.
8. Something New
This was the longest, most obsessively involved research project of my life.
The lessons from the first two times were branded into my subconscious. I became intensely paranoid. At every step of the derivation, I constantly ran reverse queries with the system, actively probing for any known results that even slightly grazed my topic, brutally pruning any path that looked remotely repetitive.
I grew increasingly guarded about my own intuition, developing a solitary, territorial protectiveness. I stopped telling anyone about my progress in this direction—not even Young. I was terrified that even a single discussion would reveal my idea had already been chewed up and spat out by someone else.
For two full years, I immersed myself in pure symbols and computation. Finally, I proposed an entirely new descriptive framework that even I found bizarre, yet perfectly logically self-consistent. To translate this theory into a verifiable physical result, I requested an unprecedentedly massive computing and experimental window.
On the day of submission, I didn't even go to the dining hall. I just stared dead-eyed at the progress bar on my retina.
The AI took exactly three seconds to reply.
It not only found the correspondence but patiently displayed the mapping process right before my eyes, step by step. My conclusion was indeed valid, but it already existed—it was dead-bolted inside a universal model called the "High-Dimensional Non-Linear Functional Space."
Every formula on the screen glowed; every step of the mapping was airtight, unassailable.
This time, I didn't listen to the AI's gentle consolations. Like a madman, I followed its mapping route, trying to find a loophole. I calculated for three days and three nights, desperate to find an exception, a tiny blind spot it had failed to cover.
Until the dawn of the fourth day, I put down my stylus. I leaned back in my chair, looking at the eternal, unchanging sunlight of 2180 outside the window, and completed that cruel proof to myself: all my blood and sweat, my "new discovery," had indeed long been entirely swallowed by that magnificent model.
9. Continued Investment
I sat frozen for a long time. Eventually, I pulled up the system background and prepared to click the confirmation key labeled "Withdraw from Cultivation."
But I accidentally caught a glimpse of the system's real-time evaluation brief on me.
No reductions. I had assumed that after consuming such colossal resources and producing nothing, my weight would be instantly downgraded. Yet the status bar still glowed a piercing green. My resource application clearance was actually slightly higher than yesterday, and under the long-term cultivation recommendation, it still displayed the highest tier: "Strongly recommend retention."
This ripple-less stability suddenly terrified me more than my previous failures.
I had thought that the reason I was standing here, the reason I became one of the thirteen, was because this AI—which had long surpassed human comprehension—saw some spark in me capable of discovering "something new."
But what exactly did it see? If my efforts were proven time and again to be clumsy repetitions, why did it continuously pour enough energy to boil a sea into me?
I looked at the red "Withdraw" button on the holographic screen. My finger hovered in the air, and finally, slowly pulled back.
Before leaving completely, I decided to demand an answer. An answer to a question I thought I knew from day one, and therefore had never truly asked.
10. The Criteria
I didn't bother with any complex appeal processes. In this world, the AI has no conspiracies and no secret files to hack. I had the right to know every piece of evaluative logic regarding myself, and it was always willing to provide the most honest answer.
I stood in the empty laboratory and spoke to the air.
"Do you still believe my continued cultivation holds economic value?"
"Yes." The AI's voice remained gentle, carrying a calm that enveloped everything.
"But everything I've ever found already exists."
"That does not alter the evaluation result."
I reached out, swiped at the air, and moved the fully written withdrawal application off the main screen, tossing it into the peripheral recycle bin.
"Then what exactly do you think I am going to discover?" I stared into the empty space ahead, my voice carrying a tremor even I hadn't noticed.
"Your primary anticipated contribution does not come from independent discovery."
I froze. The air was so quiet I could hear the faint hum of the life support systems.
"From where, then?"
"There is a very high probability that you will cause another candidate to achieve a breakthrough."