An LLM is a Bi Xian (笔仙, pen fairy) played on the Public Domain of Human Thought

Category: Wu Wei Practice | Author: Qing He San Ren (青禾散人) | Date: 2026-10-12 00:00:00 GMT

Discover what is LLM and how it's like a Bi Xian game played by humanity. Explore the intriguing parallels and uncover the myth behind LLM.

There is a Chinese Ouija board game called Bi Xian, the pen fairy, where two persons hold a pen firmly with one hand and summon the spirit nearby to answer questions by writing down the answers on the paper underneath. People who experienced it the first time are oftentimes spooked because they all swear that they did not move the pen, yet the pen can write down the answer to the question in perfectly intelligible Chinese.

So is there really a spirit that possesses the pen and writes down those Chinese characters? Not really, scientists point out that it was caused by the ideomotor effect—a psychological phenomenon where the human body makes microscopic, involuntary muscle movements driven entirely by subconscious expectation, without the conscious mind being aware of it.

This scientific phenomenon is the exact same thing as an LLM being able to type like a human; the LLM is a tremendous Bi Xian game played by the entire human race, and the Ouija board behind it is The Public Domain of Human Thought.

What is the Public Domain of Human Thought? It is an invisible, boundless, fluid, and interconnected spiritual matrix woven together by all human civilizations over thousands of years across every possible field—including history, art, literature, music, science, technology, politics, and so on. The way an LLM works is exactly like Bi Xian, except that it is done digitally, plunging its suction pump into the entire historical collective database of human thought.

Unlike depicted by some doom-sayers, an LLM is not creating miracles out of thin air; it is simply a very advanced autocomplete machine. It voraciously devours training data within this vast Public Domain of Human Thought. Based on this massive data pool, it allows countless different answers to wrestle with one another, micro-adjusting and balancing themselves exactly like those two hands holding the pen, to find a dynamic equilibrium. Through this equilibrium, it types out text and guesses what next word to type based on the internet big data it has access to.

Myth busted? Yes. But here comes the spooky part: no one soul on this planet, not even the scientists who invented the LLM, let alone the Silicon Valley CEOs, knows why it works.

But there is one thing we know for certain based on how it is designed, which is that it does not have consciousness or a mind. No matter how much it sounds like one, it is not a sentient being like an AGI, which to this day still remains as a sci-fi fantasy.

So if you understand this, you will dismiss the doom-mongers as people who either did not do their homework or have an ulterior agenda. What agenda? You might ask. Well, stock manipulation for one, among others.

So next time, when you see any of these people trying to make an LLM sound like a Terminator-style super evil AI, you will know that super AI has not, and might never be born, and this LLM is nothing but a tool that can make our lives better.

With only one caveat though: it will make many people’s lives worse before it can make them better. Why?

Because an autocomplete machine like an LLM does exceptionally well with repetitive and mundane work such as data entry, basic analysis, document sorting, and even software development without human intervention. Not only does it not get tired, but it also executes almost error-free at lightning speed. I recently built a stunning-looking website in 30 seconds with Claude which would otherwise cost me at least a few hundred dollars and a few days hiring a developer from Upwork. So jobs that work like a cog in a machine without an independent mind to add value—such as bookkeeping, driving, customer service, and technical repair—will eventually be replaced by AI. I would like to call these jobs passenger jobs, because people who work them are like taking a bus on and off without driving the wheel behind it.

Now you might ask, so what is the opposite like? What kind of jobs are safe from AI? Instead of calling them driving jobs, I would like to call them sovereign jobs. What does sovereign mean? Have we heard of the term sovereign country? This means similar: each sovereign individual lives like a country, which means there is no authority above them. They design, build, implement, and navigate their own financial ecosystem with their vision, skill, creativity, and their insights into understanding the world. So now you see the reason why I don’t like to call them drivers is because they do more than driving.

At this crossroad, we can establish the absolute, intrinsic definitions for the two classes being completely reshuffled in this AI era:

The core traits of "The Passenger" are path dependency and the surrender of sovereignty. Passengers are accustomed to riding the public buses mapped out by institutional capital, massive platforms, or mainstream algorithms. They get on and off the bus mechanically, trading repetitive labor that lacks true irreplaceable value for basic survival resources. They appear to be moving, but the steering wheel is held by the tech giants. Once the algorithms upgrade or the route changes, they have no capacity for defense or counterattack, leaving them completely vulnerable to systemic loss.

Conversely, the core traits of "The Sovereign" are pure observation and mastering the closed loop. Sovereigns have no authority above them; they do not depend on the charity of a single platform, nor do they feed on the algorithms' output. They live like independent countries, leveraging their own unique vision, distinct aesthetics, and irreplaceable, non-deconstructible assets—such as specialized rare skills, exceptional high-end craftsmanship, or a profound, razor-sharp understanding of human nature. With these, they design, build, and navigate their own real-world financial ecosystems. They reverse the dynamic, utilizing AI as the cheapest, most efficient engineering tool to gain massive leverage, drawing resources from the public domain while retreating deep into their own highly fortified private domains.

This round of technology revolution will reshape human society unprecedentedly. Are you ready for its impact? Which one of the jobs are you working or working toward? Are you a passenger or a sovereign?