Cognitive Collapse: How to Defend Our Brains in an Era of 99 Percent AI-Synthesized Data
Preface: A Detective Mystery of The More You Learn, The Dumber You Get
The story begins with a seemingly trivial incident that sent a chill down my spine.
As a veteran who has spent thirty years in the risk management department of a commercial bank, I have maintained a habit for two decades. Every Friday afternoon, stripping away all electronic devices and armed only with paper and pen, I assign two practical test cases to the young analysts on my team. I do not test them on quantitative models or complex formulas. I test them on the most fundamental, unglamorous groundwork: I hand them a company’s raw financial statements and bank transaction records, and I ask them, can we approve this loan?
In earlier years, these kids graduating from elite universities, despite their shallow experience, had solid foundations. They would circle and highlight numbers on the balance sheets with their pens. They would aggressively question why this month’s electricity bill was twenty percent lower than the previous month’s. They would dig to the very bottom to find out exactly where that abnormal accounts receivable had vanished to.
But over the past year, things have changed.
At the end of February 2026, I pulled out the background files of an abandoned real estate development project as usual. This was a real case we had handled years ago: the debtor fled, the creditor rights were labyrinthine, and it involved mortgage disputes among construction contractors, homebuyers, and multiple banks. I asked them to provide a disposal proposal.
The result?
Out of five young analysts, four handed me answers with flawless logic and crystal-clear structure, even citing the latest judicial interpretations of the Civil Code. But one glance was enough for me to know: this was the output of an AI. The phrasing was simply too standard. It was so standard it looked as if it had been photocopied from the exact same textbook.
I asked one of them, tell me, where did the actual controller of this project get his first pot of gold? Who is in his local networking circle? Is he from the same hometown as the boss of the construction firm?
The young man froze. He pushed up his glasses and whispered, sir, I used GPT to summarize the public information online, and it did not mention these details.
I replied, if the internet does not mention it, do you just stop asking? Did it never cross your mind to physically visit the project site to see what kind of neighborhood that abandoned building is in? Did you not think to dig through judicial databases to see what kind of lawsuits this boss had been involved in previously? Or even search the archived posts of local historical internet forums from ten years ago for rumors about this plot of land?
The conference room fell dead silent.
At that moment, I suddenly realized a bizarre, counter-intuitive phenomenon. In this year of 2026, where large language models are universally ubiquitous and the cost of acquiring knowledge is practically zero, many people have actually lost the capacity for independent deduction and deep penetration. Every day, they use AI to instantly summarize tens of thousands of words of research reports, watch short videos meticulously fed to them by algorithms, and listen to pundits explaining the macroeconomy with the smoothest logic. They appear to possess encyclopedic knowledge, effortlessly tossing around terms like entropy increase, closed loops, and underlying logic. Yet, the moment you drop them into the real world, facing even the smallest muddy puddle, a messy reality, or a genuine predicament fraught with information asymmetry and human frailty, their brains suddenly draw a complete blank.
What is this akin to?
It is exactly like a child raised entirely on a liquid diet, fed only pureed food. His jaw muscles and digestive tract have completely forgotten how to digest a piece of meat with bone and gristle.
This profound confusion forced me, an old-school risk manager, to dig deeper for answers. And when I opened the latest reports from the global tech community in 2026, I discovered that the crisis our human brains are undergoing is by no means an isolated case. Because in the silicon-based world, those super large models we revere as deities are undergoing an even more brutal and rapid phenomenon: cognitive autophagy.
Chapter 1: Deep Penetration: The Cognitive Trap Behind the Information Explosion
Section 1: The Data Tsunami and Forgery Crisis Triggered by Expansionary Force
Our generation was nourished by the promise of massive information. From the dawn of the internet in the 1990s, to search engines, social media, and today’s generative AI, we firmly believed in one creed: more information equals more wisdom.
But today, in 2026, this creed is being ruthlessly falsified.
The proliferation of generative AI has driven the cost of fabricating information lower than the price of scrap paper. Any middle school student can, within ten minutes, generate an industry research report that appears logically rigorous, data-rich, and properly cited. From business plans to macroeconomic analyses, from academic papers to deepfake videos, the internet is experiencing an unprecedented hyperinflation of content.
This is not a prosperity of information; it is the absolute bubbling of information.
The most severe headache for the world’s top AI scientists today is not insufficient computing power or unrefined algorithms. It is a term that sounds straight out of a sci-fi novel: Model Collapse, or AI Autophagy.
What does this mean?
Scientists discovered that when the new generation of AI large models, in an effort to cut training costs, began massively devouring texts, images, and videos generated by previous generations of AI as their learning material, something eerie occurred. The quality of the models’ outputs began to plummet. Those synthetic data generated by AI are fundamentally a low-quality replication of the real world. When these replicas are repeatedly fed as food to new AIs, errors are magnified, biases are solidified, and the original vivid, coarse, real data—full of accidents and creativity—becomes increasingly diluted.
Eventually, the model falls into a terrifying vicious cycle. It consumes its own electronic excrement and regurgitates synthetic garbage that is even more rigid, mediocre, and riddled with absurd hallucinations. This process is termed Data Pollution by the academic community.
According to international AI safety reports published in early 2026, this risk is no longer a laboratory hypothesis. With the abuse of generative AI, the proportion of synthetic data on the internet is climbing at an alarming rate. Even the Pentagon in the United States established Task Force Lima specifically to assess and defend against the information warfare risks posed by AI-generated content. Criminals are leveraging this technology, peddling packaged AI tools in underground markets, driving the threshold for cyberattacks and information fraud to an all-time low.
At this point, observant readers might ask: if those steel behemoths lying in server farms suffer from brain atrophy after consuming too much electronic garbage, then what exactly is happening to our carbon-based brains—these soft, fragile entities of flesh and blood that soak daily in AI-regurgitated articles, deepfake videos, and algorithmic echo chambers?
Section 2: Cognitive Degeneration Under Contractionary Force and the Brain’s Hedging Instinct
This brings us to the second layer of truth we must unearth today: human cognitive degeneration is, fundamentally, a passive physiological act of self-preservation.
In 2024, the Oxford English Dictionary’s Word of the Year was Brain Rot. In the academic circles of 2026, this term has become a rigorous research subject. The latest psychological and neuroscientific studies reveal that the so-called brain rot refers to the subjective decline in cognitive function and emotional blunting caused by prolonged exposure to repetitive, low-quality, algorithm-driven digital content.
Why does this occur?
We must start with the most fundamental biological instincts of the human brain. Although the brain accounts for only two percent of body weight, it consumes twenty percent of the body’s basal metabolic energy. For such an energy-intensive organ, its most instinctual reaction is not to think, but to conserve energy.
Over a long evolutionary process, the human brain developed a highly efficient cognitive miser mechanism. When faced with information, it automatically splits the processing into two systems. The fast system relies on intuition, experience, and habit, consuming almost zero energy. The slow system relies on logic, analysis, and deduction, which is massively energy-draining. On ordinary days without survival threats, the brain instinctively chooses to use the fast system to handle the vast majority of tasks, saving precious energy for unforeseen emergencies.
In the past, this mechanism protected our ancestors. But in 2026, this mechanism has become our fatal flaw.
Because at this moment, what we are facing is not information scarcity, but an information tsunami. When the infinite expansionary force—endless short videos, push notifications, and flawless AI-generated articles—surges toward us like a flood, our antiquated brain, evolved on the principle of energy conservation, triggers a lethal contractionary force reaction simply to avoid drowning.
It actively shuts down the slow system because that system is too power-hungry to withstand such high-intensity information bombardment. It begins to over-rely on the fast system, relying on the spoon-feeding of algorithms, the one-click summaries of AI, and the liquid knowledge that can be swallowed without chewing or thinking.
This is just like a person who, if fed only porridge and purees for three consecutive months, will inevitably suffer from atrophy of the jaw muscles, and their gastrointestinal tract will lose the ability to digest solid food.
By the same logic, if a young person, from high school through college for six or seven years, receives a daily diet of fifteen-second compressed videos, in-depth reports summarized by AI into five hundred bullet points, and topics precisely pushed by algorithms based on their interests, what happens when they finally graduate? When they step into a real world that requires a two-hour deep conversation to truly understand a person, months of on-site field research to penetrate a project, and a lifetime to contemplate a philosophical proposition, their brain will simply draw a blank.
They will feel tired, annoyed, and complain that they cannot understand. This is not because they are unintelligent, but because their brains have lost the essential muscle memory required for deep thinking.
Section 3: A Retrospective: PPT Dependency in Credit Approval and Cognitive Friction
In our financial circles, this cognitive degeneration actually has a more vernacular name, which existed long ago: PPT Dependency.
During the decade I oversaw credit approvals at the headquarters, I reviewed tens of thousands of enterprises. Back then, the type of client that put me on highest alert was not the companies with falsified financial statements, but the companies whose presentation decks were simply too perfect.
I recall around 2013, a cross-border e-commerce startup applied for a loan. Their financing materials were outrageously pristine. The business model canvas, the industry benchmarking analysis, the three-year cash flow projections—every chart utilized the most professional visualization tools, every data point was thoroughly referenced, and every line of copy directly struck the pain points that bank approvers cared about most. The accompanying investment banking manager, dressed in a sharp suit, was breathtakingly eloquent.
I asked the loan officer one question at the time: Have you physically visited their warehouse?
The loan officer was momentarily stunned and replied, no… their PPT states that their inventory turnover rate at the overseas warehouse is double the industry average.
I said, then we are going right now.
Without notifying the client, we drove directly to the industrial park listed as their registered address. The result? The so-called overseas warehouse was merely three rented corrugated iron sheds. The entrance was piled high with returned parcels. Because of the rain, the cardboard boxes had rotted away, and the clothes inside were covered in mold. I asked the elderly security guard, how long does it usually take to sell the goods in this warehouse? The old man took a drag of his cigarette and said, sell? The boss has not paid wages in three months. It would be a miracle if this pile of garbage could be sold.
We ultimately rejected that loan. Two years later, the company went bankrupt, owing banks two hundred million.
I remember this story to this day. Because behind it hides a profound truth about cognition.
In credit approval, the true risk is rarely the perfect numbers written on the presentation slides, but the authentic details full of friction that are hidden entirely outside the PPT.
That flawless slide deck is the perfect metaphor for AI-generated content. It is smooth, fluent, and logically self-consistent, but it is utterly frictionless. The real commercial world, however, is heavily laden with friction. It is the mud at the entrance of a warehouse, the sigh of an underpaid security guard, the abnormal fluctuation on a utility bill, and the one-second hesitation of a CEO when faced with a piercing question.
When our brains grow accustomed to frictionless information feeding, we lose the capacity to wrestle with the friction-heavy real world.
We have surrendered the defensive perimeters of our brains to machines. We believe the perfectly packaged logic, yet we no longer trust the coarse but authentic judgments derived from measuring with our own feet, observing with our own eyes, and feeling with our own hearts.
An ancient Eastern philosopher once noted that placing absolute faith in texts is worse than having no texts at all [Note 1]. If we update this wisdom for the year 2026, it should be rewritten as: To believe entirely in AI is worse than having no AI at all.
In this vast ocean of information where truth and falsehood are indistinguishable, what exactly must we rely on to anchor our cognition and avoid plunging into the abyss of cognitive collapse? This requires us to open the discussion in Chapter 2: how to construct the ultimate filter for our brains using the prism of human nature.
Chapter 2: Topological Folding: Constructing the Ultimate Filter with the Prism of Human Nature
Section 1: A Physics Perspective: The Law of Entropy Increase and Negative Entropy Assets
Let us first borrow a lens from physics to clearly view our current predicament.
The Second Law of Thermodynamics, famously known as the Law of Entropy Increase, dictates that in an isolated system without external energy intervention, the degree of chaos (entropy) will continuously increase until everything concludes in the dead silence of thermal equilibrium. This law is regarded by many physicists as the most ruthlessly profound principle in the entire scientific framework, for it points to an ultimate destiny: the irreversible arrow of time moving from order to chaos, from clarity to ambiguity.
Applied to the information world of 2026, what we are currently experiencing is an unprecedented increase in information entropy.
The homogenized content generated by AI is, by its very nature, high-entropy information. It does not originate from the complex, real-world game of survival; rather, it stems from algorithms mathematically averaging and smoothing out existing data. This content appears orderly but is essentially hollow; it seems abundant but is merely repetitive. It will not bring any new order to your brain; it will only compound the chaos and anxiety within your cognitive system.
You scroll through short videos for an hour, feeling as though you have absorbed a lot, but the moment you turn off your phone, your mind is a complete blank. This is the classic symptom of cognitive entropy increase. Your brain has absorbed a massive volume of low-quality, contradictory information fragments, yet lacks the capacity to integrate them into an ordered knowledge structure. The result is that you are more bewildered and exhausted than before you picked up the device.
So, what is the weapon to combat information entropy?
In his seminal booklet What is Life?, physicist Erwin Schrödinger proposed a deafeningly profound concept: life feeds on negative entropy. In other words, the reason a living organism can maintain its internal order and defy the law of entropy increase is by continuously extracting negative entropy from the external environment—which translates to high-quality, ordered energy and information.
In 2026, a year when even videos and voices can be perfectly forged, true knowledge and truth must possess one core characteristic: they must be negative entropy.
What is negative-entropy information? It is not those smooth, self-consistent conclusions that make you feel comfortable. It consists of first-hand facts that have undergone the coarse friction of the physical world, have been repeatedly tested by practice, and are even fraught with contradictions and flaws.
It might be the pungent smell of machine oil in a factory workshop that instantly makes you understand the concept of heavy asset depreciation. It might be the genuine price of vegetables casually spoken by a market vendor that gives you a flesh-and-blood perception of CPI statistics. It might be the drunken, unfiltered truth blurted out by an entrepreneur on the brink of ruin, allowing you to pierce through the lies of a business plan and witness the fundamental resilience and fragility of human nature.
These pieces of information possess friction. Their acquisition cost is high, they are not universally pleasing, and they cannot even be mass-produced by AI. But it is precisely this friction of acquisition that constitutes the value threshold of information. Only information that crosses this threshold earns the right to enter our brain’s slow system, to be digested, integrated, and ultimately become a structural part of our cognition, helping us reduce the chaos of the world.
As classical Eastern philosophy emphasizes, only when equilibrium and harmony are realized to the highest degree can all things attain their proper order and flourish [Note 2]. Finding this equilibrium in the ocean of information—identifying those negative entropy assets that balance reality and logic, validated by time and space—is the only way our brains can find their proper anchor, breeding truly sober cognition in an era where everything can be forged.
Section 2: Awakening the Prism of Human Nature: Anchoring Peace of Mind in the Torrent of Algorithms
At this juncture, a thornier problem emerges: even if we know we must pursue negative entropy, how on earth do we filter it from the daily deluge of hundreds of millions of data points? What criteria allow us to instantaneously judge whether a piece of information is an ore worth mining or slag that should be immediately discarded?
This brings us back to a core concept we have repeatedly emphasized: the Prism of Human Nature.
AI possesses no values. This is the starting point of all our discussions. It can, upon your command, generate ten thousand different business models, ranging from blockchain pig farming to metaverse funerals, each with a logical chain more rigorous than the last. But it can never answer the most fundamental question for you: which choice will allow your inner self to find true peace of mind?
Home is where the heart finds peace [Note 3]. This poetic sentiment acquires profound contemporary relevance in the information explosion of 2026. It is no longer just about a geographic hometown; it is about a spiritual anchor.
Deep within each of us lies a unique, non-economic, stable belief system rooted in value rationality. This is your Prism of Human Nature. It might be a cherishing of love and belonging—such as choosing to abandon a high-paying job in a tier-one city to return to your parents, even if it means a dimensional downgrade in career development. It might be a nostalgia for local culture—like an old artisan persisting in a declining county town, using manual crafts to resist the wave of industrialization. It might be an obsession with a certain lifestyle—such as rejecting the grueling 996 work culture, preferring to earn less just to pick up your child from school at four o’clock every afternoon. Or it could be unshakeable ethics or faith—like walking away from a highly lucrative industry because it crosses your moral bottom line.
In the eyes of economists, this belief system might be deemed irrational because it does not conform to the assumption of profit maximization. But in our framework, it is precisely the most noble and irreplaceable part of a human being that no algorithm can ever simulate.
In an era flooded with information, your internal moral bottom lines, your local sentiments, your non-negotiable life boundaries—which constitute your unique Prism of Human Nature—serve as the sole value anchor for filtering massive amounts of data.
This statement bears repeating three times.
Why?
Because when AI can generate any logic, logical self-consistency itself loses its screening value. A meticulously fabricated lie can be more logical and perfect than the truth. At that moment, what helps you make a judgment is not your proud rationality, but your seemingly irrational value intuition.
Information that does not align with your prism’s values, no matter how gorgeously packaged or rigorously argued, is merely noise to you—even toxic noise. It will disrupt your peace of mind, drain your energy, and cause you to gradually lose yourself while chasing a definition of success authored by someone else.
Conversely, information that aligns with your prism’s values—even if it is just a coarse clue, a microscopic opportunity, or a seemingly unprofitable choice—will instinctively evoke a sense of being summoned within you. You will be willing to invest time, energy, and even emotion into it. And this posture of willingness to invest is exactly the starting point where deep cognition occurs.
Ancient foundational texts speak of illustrating illustrious virtue and resting in the highest excellence [Note 4]. This resting is about anchoring, knowing where your boundaries lie, and knowing exactly where you should stand. In the vast ocean of information, your Prism of Human Nature is the solid rock beneath your feet. With it, you will not be swept away by the tide; you can stand firm and see your direction clearly.
Section 3: Evolutionarily Stable Strategy: The Authentic Sense of Pain is the Most Expensive Moat
From a biological perspective, the existence of this Prism of Human Nature carries an even deeper evolutionary significance.
In the 1970s, John Maynard Smith introduced a profoundly influential concept: the Evolutionarily Stable Strategy (ESS). Simply put, in a population, once the majority of individuals adopt a certain strategy, any minority deviating from this strategy cannot gain a greater survival advantage. This strategy thus stabilizes through evolution.
Applying this to the evolution of the information environment: when the entire ecosystem is saturated with forged signals, synthetic data, and deepfake content, what becomes the most stable survival strategy?
The answer is: returning to simplicity, reverting to authentic signals that absolutely cannot be forged.
In biology, there is a classic example: the peacock’s tail. That heavy, flamboyant tail is actually a burden in the survival competition, easily attracting predators. Yet it has become an unforgeable honest signal—because only the genetically superior and physically strongest males can afford to bear this burden, possessing the capital to flaunt their advantage in such a wasteful manner. Weak males could not forge it even if they tried.
In the information world of 2026, what signals are unforgeable?
It is the authentic, instinctual reactions generated by your physical body after direct contact with the physical world.
You go to a factory workshop and smell the machine oil. If it is pungent, your body will cough, and your eyes will water. This physiological discomfort is real; AI cannot forge it.
You go to the wet market and ask a vendor the price of vegetables today. The number she casually tosses out is backed by the exhaustion of waking up at 3 AM to stock up, representing the most authentic fluctuation of supply and demand in the city. This number carries body temperature; AI cannot forge it.
You go drink with that entrepreneur on the brink of bankruptcy, listen to his drunken tears, see the bloodshot veins in his eyes, and feel the desperate yet unyielding grip of his handshake. This microscopic glimmer of humanity in despair is something no script could ever write, and AI certainly cannot forge it.
The authentic sense of pain, the authentic exhaustion, the authentic joy, the authentic fear—these first-hand experiences smelling of the earth constitute the most foundational Evolutionarily Stable Strategy of human cognition.
In my thirty years at the bank, the people who commanded my deepest reverence were not the financial engineers sitting in CBD office towers calculating infinite possibilities with models. They were the grassroots loan officers who spent three hundred and sixty-five days a year immersed in the fields, knowing the financial bottom lines and temperaments of every single farming household inside out. They might not understand complex financial derivatives, but they possessed one superpower: no fraudster could last more than three sentences in front of them.
Why? Because their brains stored massive volumes of samples derived directly from the real world. No matter how flawless a scammer’s rhetoric was, if just one detail failed to align with a real case they had encountered, their intuition would instantly sound the alarm. This intuition is not innate talent; it is a moat built inch by inch by measuring the world with the soles of their feet.
An ancient philosopher once taught that if a person’s words are sincere and truthful, and their actions honorable and careful, they will succeed even in the most barbaric lands; but if they lack these traits, they will find it impossible to take a single step even in their own hometown [Note 5].
This wisdom finds a new annotation in 2026. In an era where the boundary between the virtual and the real is increasingly blurred, what we ultimately rely upon is not superb debating skills or flashy mental models, but the plainest, yet most expensive mandate: words sincere and truthful, actions honorable and careful. Use real footsteps to measure the real world; use a sincere heart to perceive real human nature. This is our final, most impregnable line of defense as carbon-based life forms against the silicon-based flood.
Chapter 3: Phase Transition Leap: Becoming a Cognitive Architect to Master the Silicon Flood
Section 1: Proactively Creating Cognitive Friction: Squeezing the Last Drop of Wisdom from AI with the Five-Dimensional Phase Transition Law
Information overload and AI synthetic pollution are plunging our cognitive systems into terrifying entropy increase; our only anchor is the Prism of Human Nature, which algorithms cannot replicate. But the subsequent problem arises: in an era where AI is deeply embedded in workflows, we cannot completely reject technology and revert to the era of bamboo scrolls. So, how exactly do we coexist with technology? How do we maintain the sharpness of our brains while surrounded by AI?
The answer is: we must upgrade from being passive information consumers to becoming proactive cognitive architects. And the core tool to achieve this leap is the Five-Dimensional Phase Transition Law we have repeatedly honed.
Many people use AI with a search mindset: I want a report on the new energy industry, generate one for me. This is an outsourcing mindset. You outsource the labor of thinking to the machine, and what you get back is an average answer cobbled together by the machine from massive synthetic data. This answer is smooth, fluent, and entirely safe, but it is utterly worthless—because it has not undergone any cognitive friction and has not sparked a single intellectual flame.
True masters utilize a phase transition mindset. They treat AI as a colossal, cross-disciplinary knowledge reactor. They proactively inject thinking catalysts from various disciplines into it, forcing it to undergo a phase transition amidst the violent collision of heterogeneous knowledge—yielding counter-intuitive new concepts and models with immense explanatory power.
Let me demonstrate what it means to proactively create cognitive friction.
Suppose today we need to research a classic economic problem: why do many small and medium-sized enterprises (SMEs) refuse to take out loans despite the loose credit environment in 2026? If handed to a standard AI, it will list the textbook answers: low expected returns, declining risk appetite, banks’ reluctance to lend. These are all correct, but they do not transcend anyone’s cognitive boundaries.
But if we activate the Five-Dimensional Phase Transition Law, the situation completely changes.
Dimension One: Breadth—Panoramic Scanning. I command the AI: Please analyze the evolution of bank-enterprise relationships through the biological concepts of symbiosis and parasitism. The AI retrieves: in nature, a symbiotic relationship requires continuous contact and signal exchange to establish trust; once the environment abruptly changes, one party may transition to parasitism, draining the host’s nutrients until it dies. This makes me wonder: is the current relationship between banks and enterprises sliding from symbiosis to parasitism? Banks force loans to meet inclusive finance quotas; enterprises take the money not to expand production, but to deposit it in wealth management products to earn the interest spread. Is this not a form of parasitism?
Dimension Two: Depth—Principle Penetration. I probe further: Please use the physics principle of entropy increase to explain why credit easing actually leads to capital stagnation. The AI tells you: when energy (currency) is forcefully injected into a system that lacks an effective dissipative structure (real investment opportunities), the energy dissipates in a disordered form, manifesting as capital idling and asset price bubbles. This shows me that credit easing is not the problem; the lack of a dissipative structure is the root cause.
Dimension Three: Correlation—Topological Folding. I throw an even wilder prompt: Please attempt to reconstruct the trust relationship between banks and enterprises using the concept of polyphony from music theory. The AI might respond: in polyphonic music, multiple independent voices proceed simultaneously, maintaining their individuality while forming a harmonious whole. This serves as a metaphor for the ideal bank-enterprise relationship. Banks and enterprises should be two independent voices, each with its own melody (business logic), forming a complex harmony under the government’s conducting. In reality, we often force enterprises to abandon their own melody to cater to the bank’s main theme, resulting in monotonous music.
Dimension Four: Time—Evolutionary Deduction. I ask the AI to trace history: Please review the creation of ‘Jiaozi’ (early paper money) in the Northern Song Dynasty and how it collapsed after over-issuance, and use this to deduce the potential risks of today’s digital currency. The AI takes you back: Jiaozi began as private credit, was later nationalized, and led to hyperinflation due to infinite over-issuance. This warns us that the evolution of any currency form cannot escape the erosion of the balancing force by human greed.
Dimension Five: Core—Value Anchoring. Finally, I ask the AI to return to the essence: What does all this discussion ultimately point to? For that SME owner agonizing late at night, what is true negative entropy? The AI might fall silent, because it has no values. But the question itself forces me to think: the true solution is not more loans, but helping him find the business direction that allows him to find his peace of mind.
You see, through the forced collision across these five dimensions, what I extract from the AI is not a standard answer, but a pile of intellectual ores—the metaphor of symbiosis and parasitism, the revelation of dissipative structures, the harmony of polyphony, the lesson of the Jiaozi collapse. These fragments require my own brain to chew, integrate, and smelt, ultimately potentially birthing a brand-new cognitive model. This process is cognitive friction. It is mentally draining and exhausting, but it is the only valuable thing I can squeeze out of AI.
Ancient scholars advised to study extensively, inquire carefully, think thoroughly, distinguish clearly, and practice earnestly [Note 6]. In 2026, studying extensively has been outsourced to AI; it can read all of human literature in a single day. But inquiring carefully, thinking thoroughly, distinguishing clearly, and practicing earnestly—these steps that require friction—must be executed by us. AI is the miner of knowledge, but the sorting, smelting, and forging are our own responsibilities.
Section 2: Downgrading AI to a Tool and Upgrading the Self to a Judge: Establishing Your First-Principles Tribunal
Proactively creating cognitive friction solves the problem of how to use AI. But the deeper challenge lies here: when we are bombarded daily by countless contradictory information and opinions, who makes the final call? Who acts as the judge?
Many people operate like this: whoever speaks sounds reasonable, and ultimately they listen to whoever is trending. Today, Professor Zhang says inflation is coming, so they rush to hoard gold. Tomorrow, Dr. Li says deflation is the main theme, so they scramble to sell gold for cash. In this state, a person becomes a slave to information, led by the nose by various viewpoints, while their anxiety mounts daily.
To escape this state, a fundamental transfer of power must occur. You must seize the power to filter information and make final decisions away from algorithms and experts, and hand it back to yourself.
How do you seize it? You need to establish your own First Principles common sense judgment framework.
First principles is a concept from physics and philosophy, referring to fundamental propositions that are indivisible and self-evident. For instance, in physics, energy conservation is a first principle; in ethics, do not do unto others what you do not want done to yourself can be considered a first principle. Any complex theory must withstand the test of first principles; any conclusion conflicting with reality must be corrected by first principles.
For ordinary people, establishing a first principles framework does not require profound academic knowledge. It is actually about distilling the underlying common sense you have acquired and repeatedly verified in the real world into a checklist. Whenever you encounter new information or opinions, you pull out this checklist and test them item by item.
For example, as a veteran who spent thirty years in banking, my brain holds many such underlying common sense principles:
Principle One: Any business model must ultimately prove itself through cash flow. If a business plan boasts extravagant claims but cannot calculate when positive cash flow will appear, it is pure charlatanism.
Principle Two: Human nature is unreliable in the face of profit. Any trust lacking institutional constraints is a ticking time bomb.
Principle Three: High returns are always accompanied by high risks. If someone pitches you a low-risk, high-return opportunity, they are either a fool or a fraud.
Principle Four: The development of things is tortuous. Any prediction claiming a straight-line ascent or inevitable success is a lie.
Principle Five: The opportunities seen by the vast majority are actually traps; true opportunities often hide where the majority cannot see, look down upon, or cannot understand.
This set of common sense principles is not textbook theory. It is something I ground out inch by inch over thirty years—from piles of bad debts, recurring crises, the tears of failures, and the cold sweat of successes. It may not be delicate, but it is extremely robust.
Now, facing any AI-generated viewpoint, expert assertion, or trending internet topic, I pull out this set of common sense principles as the constitution of my tribunal and put them on trial.
For instance, if an AI model predicts a certain stock is guaranteed to rise next year because technical indicators look good and capital is flowing in, I bring out Principle One: How is its cash flow? If it does not even have proper financial statements, on what grounds should I believe it? Overruled.
If a trending economist screams, Global liquidity is surging, hurry up and buy real estate! I bring out Principle Four: Will housing prices rise in a straight line? Do the fundamentals—urbanization rates, demographic structure, household leverage—support this? Held in doubt.
If a friend recommends a guaranteed no-loss wealth management project with a 15% annualized return, I bring out Principle Three: What are the underlying assets? Why is it offering such high returns? If the other party stammers, instantly blocked.
This First-Principles Tribunal does not require starting from scratch with lengthy arguments every time, nor does it require digging through mountains of literature. It acts as an automatic filter, instantaneously screening out 99 percent of garbage information and toxic opinions. Only the remaining one percent is worth my time and energy to initiate the five-dimensional phase transition for deep cognitive processing.
Classical military strategy states: Make oneself invincible first, then wait for the enemy’s vulnerability [Note 7]. On the battlefield of information, establishing a first-principles framework is precisely making oneself invincible first. I fortify my cognitive defense lines so that fake, misleading, and toxic information collapses upon contact. Only then can I calmly wait for truly valuable information to appear, capturing the opportunities that allow for cognitive leaps.
Section 3: Conclusion and Reshaping: Your Altitude Depends on How Much Garbage You Discard
Let us return to the starting point of this article.
That young person in 2026—surrounded by AI, spoon-fed by algorithms, and drowning in synthetic data—felt genuine bewilderment and exhaustion. His brain is indeed undergoing an unprecedented cognitive collapse. But this crisis is not without an antidote.
The antidote is not outside; it resides within our own bodies.
It is our brain, shaped by millions of years of evolution, inherently equipped with an energy-saving instinct. We must overcome its inertia and reactivate that forgotten slow system.
It is our unique Prism of Human Nature. We must awaken it and use it as the most foundational value anchor to stand firm amidst the information torrent.
It is our feet capable of measuring the real world, our hands capable of touching real objects, and our hearts capable of feeling real sorrow and joy. These authentic signals from the physical world are negative entropy assets that no AI can ever forge.
And it is our First Principles common sense honed through practice. This internalized cognitive framework is our most reliable information filter.
By combining these elements, you become a true Cognitive Architect. You are no longer a passive vessel receiving information, but an engineer actively constructing cognition. You know when to use AI and when to trust yourself; you know which information merits deep digging and which should go straight into the trash bin.
And in this process, you will increasingly recognize a counter-intuitive truth: your altitude is determined not by how much information you acquire, but by how decisively you discard the garbage.
Ancient Eastern philosophy teaches that in the pursuit of learning, every day something is acquired; in the pursuit of the Way, every day something is dropped. Drop and drop again, until you reach non-action [Note 8]. Pursuing knowledge requires daily addition, but pursuing the ultimate truth requires daily subtraction. Subtracting the illusory, the superfluous, and the mind-disturbing noise, until there is nothing left to subtract—only then can you touch the most essential truth.
In the information world of 2026, daily acquisition has been delegated to AI; it can read all human literature in a day. But daily dropping must be executed by ourselves. Drop the synthetic nonsense generated by AI, drop the emotional garbage fed by algorithms, drop the noise that contradicts our prism’s values. Drop and drop again, until our brains are left with only the truly vital things that have withstood the test of the real world and resonate with our very lives.
Only then do you truly possess the wisdom to survive and thrive in this era
Conclusion: Guarding Carbon-Based Clarity in a Silicon Illusion
Frankly speaking, as I write these words, my heart is not light.
During my thirty years at the bank, I lived through the Asian Financial Crisis, the 2008 global meltdown, and countless market booms and busts. But never has any event made me as profoundly anxious about our future cognitive environment as I am today. Because this time, the crisis does not originate externally, but from within our own brains. We are personally constructing an information echo chamber, locking ourselves inside, and using algorithms to feed ourselves synthetic food.
I know that you, reading this article, might have also aimlessly scrolled through your phone late at night, feeling an inexplicable emptiness. You might have been swept up by contradictory information, not knowing who to believe. You might have doubted whether you could still maintain your clarity in this era where truth and fiction are indistinguishable.
Your bewilderment, exhaustion, and anxiety are real and deserve to be understood. Because this is the shared destiny of our generation.
But it is precisely this struggle of panning for gold in the mud, this obsession with parting the fog to see the truth of the world, this stubborn refusal to drift with the tide or be domesticated by algorithms, that proves our most noble value as human beings.
AI can generate perfect articles, but it cannot comprehend why your eyes well up with tears when you read a certain sentence.
AI can simulate affectionate tones, but it cannot feel why your heart races when you hear a certain story.
AI can output infinite knowledge, but it cannot know the profound meaning of the euphoric clarity you experience when you finally figure out a problem.
These are the irreplaceable attributes of our existence as carbon-based life forms. They are also our most impregnable fortress against the silicon-based flood.
Finally, allow me to conclude this article with a timeless quote from ancient wisdom:
Learning without thought is labor lost; thought without learning is perilous [Note 9].
In 2026, this quote has acquired a brutally contemporary footnote. Merely absorbing AI-generated data without independent thought leads to a cognitive collapse akin to model collapse. Conversely, purely relying on one’s own isolated pondering without leveraging AI to gain broad perspectives leads to the danger of narrow-mindedness and bigotry. Only by finding that fragile equilibrium between learning and thinking, between AI and the Self, can we guard our brains and preserve the clarity and dignity belonging to humanity in this massive cognitive reshuffle.
If you are willing to walk alongside us in this protracted war for cognitive defense, you are welcome to follow the Wise Companions community: Finsages. Here, my book manuscripts Cognitive Architect and The Sword of the Architect await, panning the true gold and silver of this era for you.
Walk with the wise, and seek the optimal solution amidst uncertainty.
【注释合集 / Footnotes】
[Note 1] This reflects the sentiment of Mencius: “尽信《书》，则不如无《书》” (To believe entirely in a book is worse than having no book at all). It serves as a profound warning against blind faith in authoritative texts or, in a modern context, synthetic data.[Note 2] An adaptation from the Doctrine of the Mean (Zhongyong): “致中和，天地位焉，万物育焉” (When equilibrium and harmony are realized to the highest degree, heaven and earth will attain their proper order and all things will flourish). It emphasizes the crucial role of balanced, grounded truths in maintaining systemic order.
[Note 3] A famous line from the Song Dynasty poet Su Shi: “此心安处是吾乡” (Home is where the heart finds peace). It highlights the psychological anchor derived from one’s core value system.
[Note 4] Core concepts from the Great Learning (Daxue): “明明德” (to illustrate illustrious virtue) and “止于至善” (to rest in the highest excellence), underscoring the necessity of knowing one’s moral and philosophical boundaries.
[Note 5] From the Analects of Confucius (Lunyu): “言忠信，行笃敬，虽蛮貊之邦，行矣。言不忠信，行不笃敬，虽州里，行乎哉？” (If a man’s words are sincere and truthful, and his actions honorable and careful, he will get along even among barbarians. If not, how can he get along even in his own neighborhood?). It establishes authenticity as the ultimate, unforgeable survival strategy.
[Note 6] An epistemological framework from the Doctrine of the Mean: “博学之，审问之，慎思之，明辨之，笃行之” (Study extensively, inquire carefully, think thoroughly, distinguish clearly, and practice earnestly).
[Note 7] A strategic principle from The Art of War by Sun Tzu: “先为不可胜，以待敌之可胜” (Make oneself invincible first, then wait for the enemy’s vulnerability).
[Note 8] A philosophical observation from Laozi’s Tao Te Ching: “为学日益，为道日损。损之又损，以至于无为” (In the pursuit of learning, every day something is acquired. In the pursuit of the Tao, every day something is dropped. Drop and drop again, until non-action is reached).
[Note 9] A classic educational maxim from the Analects of Confucius: “学而不思则罔，思而不学则殆” (Learning without thought is labor lost; thought without learning is perilous).
