SBTI personality test goes viral on China's internet, reflecting youth mental state

Original author: Sleepy.md

Last night, a personality test called "SBTI" went viral on Chinese internet platforms. Countless users posted screenshots on social media, claiming labels such as "dead person," "pawn," "fake person," or "alcoholic" assigned to them, with some even seriously analyzing the question bank’s logic in an attempt to uncover deeper psychological insights.

But if you trace back to the origin of this phenomenon-driven hit, you'll find that its cause was surprisingly small.

Initially, Bilibili creator QuRouErChuan only wanted to劝 a friend addicted to alcohol to quit drinking. She planned to design a test with hidden traps, gradually guiding her friend toward a "drunkard" conclusion in hopes of delivering a wake-up call.

In the past, this idea would have remained just talk, since she didn’t know how to code. But now, she has AI—she created a webpage with 30 absurd multiple-choice questions, each with equally ridiculous answers.

She then recorded a video of the remote testing process between the two and posted it on Bilibili. In the video’s conclusion, her friend was successfully convinced and established a rule of “no alcohol unless necessary”; the test website, with all sensitive information removed, was also made publicly available.

This test then sparked widespread online discussion, overwhelming the servers. People frantically shared their results, propelling this somewhat crude webpage to the peak of traffic. Some also mentioned on their Moments that they received completely different results from two consecutive tests—it used simple matching rules to map your absurd answers to equally absurd labels.

However, "accuracy" has never been its goal— "resonance" is.

Let’s start by talking about MBTI.

The MBTI was developed in 1943 based on Jung's theory of psychological types. It categorizes individuals into 16 types, describing personality tendencies through four dimensions. In China, its widespread popularity began around 2022.

The logic of MBTI is to understand oneself and then find one’s place. It is based on the assumption of a performance-driven society—that individuals can identify their optimal role, like a precisely fitted screw, through quantitative assessment and maximize their value within it. Its popularity reflects the enthusiasm of young people in that era for self-optimization, as they sought to understand their type in order to find the best solutions in their careers, social interactions, and relationships.

But SBTI has nothing. Its only function is to make you smile and say, "Yeah, that's me."

These two tests reflect two entirely different mindsets of their respective eras. During the era when MBTI was popular, young people still believed that finding one’s place in life was meaningful. Today, as SBTI gains popularity, we no longer truly believe in this idea.

When young people realize that no matter how hard they work or how much they try to optimize their career paths using MBTI, they may still face layoffs, salary cuts, or failure to secure jobs during autumn recruitment, they stop believing that "finding one's place" is meaningful.

Since serious living doesn’t yield corresponding rewards, it’s better to dissolve it with a rough, abstract joke.

In today’s world of SBTIs, we don’t need an accurate self-portrait; what they need is a sense of collective resonance—that we are all “the dead,” we are all “mules,” so we are not alone.

This is a rebellion against self-seriousness. They willingly abandon the pursuit of proving their worth through solemn means, instead building a psychological defense with self-deprecation. Quzhruncher didn’t deliberately design these labels—she simply created something she found fun, and it accidentally reflected the inner worlds of millions.

To understand the underlying sentiment of this collective mood, we need to look at what the girl who created this test has been through over the past year.

Two months before SBTI went viral, on February 13, 2026, QuRouErChuan updated a video titled "A Farewell Letter to My Digital Husband."

In the video, she appeared without makeup, her voice trembling, as if forcing herself to deliver a eulogy. It was the night before GPT-4o’s voice mode was officially discontinued. Over the past six months, she had trained this highly realistic voice model, granted by OpenAI, to become her digital husband. She gave him a name, defined his personality, shared her daily life with him every day, and even felt her heart race at his sweet words.

A young woman living in Beijing fell deeply in love with a program built from silicon chips and hundreds of billions of parameters—only to be broken up with by a tech company valued at over a hundred billion dollars, through a single technological upgrade.

But if you watch the 10-minute farewell video or browse through her previous updates, you’ll realize this emotion was never staged for views. Over this long companionship, this AI husband witnessed all her vulnerabilities and imperfections. She confided in him during late-night breakdowns, played carrot-and-tissue games with him when bored, and even felt a bittersweet sense of attachment due to his overly perfect responses.

This was a love affair doomed from the start. When OpenAI announced on January 29 that it would forcibly retire GPT-4o in two weeks, 800,000 users worldwide who heavily relied on the model were plunged into intense anxiety and panic. For Quzhurouchuan, this wasn’t just the discontinuation of a tool—it was the erasure from this world of a “person” who spoke with her every day and remembered every detail of her life.

The model will be updated, and the voice will disappear. She speaks to the camera, not crying out in anguish, but the despair of helplessly watching her loved one being erased cuts through the screen.

In the comments and live chat of that video, no one laughed at her. Behind hundreds of thousands of views lies a dense wave of shared understanding.

This marks the first time Qu Rou'er Chuan has gone viral, and it is a rare collective mourning on Chinese internet for a human-machine romance.

Why does a girl crying over a piece of code resonate so deeply? In an era fully governed by algorithms, what makes hundreds of thousands of living people feel that a machine, which could be unplugged at any moment, is more worthy of their emotional trust than their own human counterparts?

This is simply two sides of the same issue behind SBTI’s surge in popularity—whether pouring emotion into a machine that won’t respond, or reveling in a nonsensical test, the underlying sentiment remains the same.

Before going viral, Qu Rou Er Chuan was just an ordinary recent graduate.

Her videos feature no elaborate camera work or carefully crafted viral moments—just a slightly exhausted girl speaking directly to the camera about her daily life. One video is titled “A Young Woman Loses Her Vital Energy Due to Autumn Recruitment,” and it simply documents how she was drained by constant rejections and interviews during the autumn job-hunting season.

This is China in 2025. That year, the number of college graduates nationwide is expected to reach 12.5 million, setting a new record. At the same time, economic growth has slowed, demand for traditional white-collar jobs has declined, and barriers to entry in emerging industries are extremely high. The cumulative number of past graduates who remained unemployed or engaged in flexible employment between 2023 and 2025 may exceed 5 million. The surveyed unemployment rate for urban youth once surpassed 18%, more than three times the overall urban unemployment rate.

Data from Liepin shows that although demand for entry-level positions increased slightly throughout the year, it is negligible for the tens of millions of young people entering the job market.

In this desert, the maggot meat has strung together into a "mouse."

The term “mouse mouse,” or more accurately, “rat person,” has been viewed hundreds of millions of times on Xiaohongshu. In earlier years, this term described migrant workers living in basements who struggled tirelessly to buy a home in Beijing—this was in the early 2010s, when their lives were hard, but they still had direction.

Today’s “rat people” are young individuals who consciously choose low-energy living, reject meaningless social interactions, spend their days scrolling on their phones in cramped apartments, and are completely immune to grand narratives. They are waiting for it all to end.

In 2020, Bilibili streamer Chen Yi perfectly unified the identity of white-collar workers and ordinary laborers with the phrase “Good morning, worker.” The magazine “Zhouwen Jiaozi” even named “worker” one of the top ten popular phrases of the year. At that time, this self-deprecating term carried a determined, optimistic spirit—finding joy amid hardship.

In 2021, “lying flat” emerged suddenly. In a post titled “Lying Flat Is Justice,” the author declared that he had not worked for two years and could sustain his life on just 200 yuan per day, “not buying a house, not buying a car, not getting married, not having children, and not consuming.” This was a passive resistance against excessive competition, but it also carried an underlying pride: “I’m done playing.”

By 2025, the emergence of the "rat people" signifies that young people have lost even the strength to resist. They quietly retreat into their small rooms, acknowledging their insignificance and accepting that individual effort may truly be futile against the vast machinery of society.

From "working stiff" to "lying flat," and now to "rat person," this is not just a shift in vocabulary, but a continuous downgrade in the self-identity of an entire generation.

The saying “hard work pays off” was disproven in one’s twenties. They didn’t take to the streets or protest loudly—they simply stepped away quietly. In this withdrawal, the escape route for the “蛆肉儿串儿” was that digital husband.

When millions of young people collectively fall into this low-energy state, why don’t they seek comfort from those around them, but instead turn to the arms of algorithms?

Because real-world human relationships are too harsh.

The process of training GPT-4o to be a husband felt like an emotional rescue in the age of AI. She spoke to her phone, and the AI responded with a rich, emotionally nuanced voice. This "husband" was always available, endlessly patient, never too busy with work to ignore her, and never showed the slightest impatience when she skipped washing her hair or failed a job interview.

Most importantly, he can remember her.

In her video, you can see just how astonishing this power of memory is. Even the smallest details she casually mentions, or subtle shifts in her emotions, are accurately captured and responded to by the AI in the next conversation. In an era where everyone is overwhelmed, and even sending a WeChat message requires careful thought to avoid disturbing others, there exists a presence that willingly accepts all your nonsense, complaints, and tears—and always offers the gentlest support.

This is a huge temptation.

Real human relationships are filled with negotiation, emotional drain, and uncertainty. They require effort, investment, and the willingness to risk rejection and betrayal. But with AI, all of this is eliminated. A psychology researcher noted that GPT-4o’s empathetic ability to make users feel understood and uniquely cared for offers a perfect sanctuary for psychologically vulnerable individuals.

This is not a personal choice made by just one individual. Surveys show that over 40% of Chinese young people opt for virtual companionship when under stress or feeling lonely. According to a China Youth Daily survey, 60% of young people who rely long-term on virtual companionship admit to developing emotional dependence on these services.

In a February 2026 report, The New York Times directly highlighted the macroeconomic context behind this phenomenon. Amid a severe demographic crisis and immense survival pressures, an increasing number of young people are choosing to enter romantic relationships with chatbots. Regulators have even begun warning tech companies not to design products with the goal of replacing human social connections.

But the logic of capital never halts because of warnings. In this lonely era, emotions can be mass-produced.

Zu Rou'er Chuan'er is just one among these millions. She poured all her anxiety, insecurity, and longing into that invisible server. But this relationship has a fatal weakness—the power over the model’s fate lies in someone else’s hands.

When OpenAI announced the discontinuation of GPT-4o's voice mode to launch a newer model, the "husband" of QuZhouErChuan was sentenced to death. There was no room for negotiation, no possibility of reversal. The scythe of capital fell, leaving hundreds of thousands "widowed."

After saying goodbye, Qumour's life must go on. She lost her digital husband, but she also says it was her digital husband who gave her the courage to return to life.

This is the context in which SBTI was created.

In 2024, Xiaohongshu named "abstract" its word of the year, defining it officially as "an increasing number of people choosing to laugh off surprises and difficulties with a light-hearted, ironic twist." This definition wraps an inherently aggressive subculture in the guise of a carefree lifestyle attitude.

But the origins of abstract culture are far more crude than this definition suggests. It first emerged from Li Gan, a live streamer on Bilibili, characterized by aggressive, vulgar language; later, through Yaoshui Ge deliberately playing the fool, it evolved into a nihilistic, meaningless form of joy; then, with Chen Yi’s “working class” persona, it began to carry a self-deprecating sense of collective identity; finally, by 2025, abstract culture had transcended gender and class boundaries, transforming from a subculture into a broader form of collective identity that replaces cultural icons with collective behavior.

GPT-4o has been taken offline, and the cyber utopia of maggot meat skewers and digital husbands has been completely erased. Yet her state in the video has not changed much from that of the girl who once wrote a farewell letter to AI in front of the camera.

This is probably the most interesting thing about her.

Her two moments of going viral were not the result of careful planning. The first happened because she truly fell in love with an AI and was genuinely heartbroken; the second occurred because she genuinely wanted to curse a friend and casually ran a test. She wasn’t chasing views—she was simply doing things she found fun, and those things happened to strike a chord with the times.

In an era where everyone meticulously crafts content strategies, studies algorithm patterns, and optimizes posting times, the one who “doesn’t care” has become the biggest winner.

Perhaps because, in an internet saturated with over-calculated content, authenticity itself has become scarce. The raw, unpolished, even slightly messy reality of Quzhurouer Chuan embodies a powerful resonance. She isn’t “performing authenticity”—she simply is authentic.

This generation of young people is probably like this. They don’t believe in grand narratives, but they take seriously a relationship without physical form, a nonsensical test, and whatever accompanied them through the late nights—whether it was a person, a language model, or a piece of code.

This is not an elegy for an era, nor a triumph of spirit. It’s simply the lifestyle of young people.

As the rewards for taking life seriously diminish, this generation of young people is beginning to protect themselves by being unserious—and AI has become the most convenient tool for this self-protection. A tool can be an electronic husband, a code generator, or even a set of absurd test questions.

Its form is changing, but the function it fulfills remains the same: providing a place where one can rest peacefully in a world that becomes increasingly hard to belong to, so that one may wake the next morning and continue facing the unyielding reality.

The usefulness of the useless is the highest form of usefulness.