E Fund Chairman Liu Xiaoyan Appointed as Part-Time President of AMAC's New Council
Monthly active users evaporated by 6.1 million overnight. Doubao's charge came as a heavy blow, striking itself squarely in the face. But this is far from just one company's embarrassment—it's an iceberg that the entire AI-native app sector has collectively collided with. Users' expectations of a "free lunch" and companies' thirst for a "cash cow" have finally clashed.
Analysis
Monthly active users evaporated by 6.1 million overnight. Doubao's charge came as a heavy blow, striking itself squarely in the face. But this is far from just one company's embarrassment—it's an iceberg that the entire AI-native app sector has collectively collided with. Users' expectations of a "free lunch" and companies' thirst for a "cash cow" have finally clashed.
Over the past two years, all major model providers have been aggressively expanding their territories using "free" and "subsidy" strategies, as if user growth were the sole articles of faith. Metrics like monthly active users, daily active users, and download counts—relics of the internet era—have been elevated once again to sacred status. Until the bills arrived: computing costs were a bottomless black hole, ad monetization remained a distant prospect, and investor patience was drying up alongside valuation bubbles. Charging became an inevitable "coming-of-age" ritual. However, users voted with their feet far faster than expected. The 6.1-million-user exodus starkly reveals a truth: for most people, their "need" for AI is still confined to superficial levels of "curiosity" and "fun." The moment it shifts from free to paid, they immediately calculate the cost-benefit ratio—does this productivity boost justify dozens of yuan? The answer is clearly no. This is not Doubao's failure, but a severe misjudgment by the entire industry regarding "user stickiness." We overestimated the depth of AI's integration into daily life and underestimated the human obsession with "free."
Just as providers struggle with commercialization, another warning emerges from the deeper waters of technological ethics. Anthropic's statement urging a global slowdown in AI development, with the phrase "self-improvement" in its title, carries the chill of a sci-fi thriller. This is no longer a distant concern but a whistle blown at the industry's current accelerated intensification. Researchers in labs are pursuing absolute intelligence in models while fearing the emergence of uncontrollable autonomy. This contradiction resembles the myth of a child repeatedly dropping stones into an abyss to measure its depth. Once the Pandora's box of "self-improvement" is opened, will humanity gain an omniscient assistant, or will we unwittingly nurture an incomprehensible "alien"? Anthropic's warning is less a moral appeal than a shiver of survival instinct—they know best that what they are creating, with its complexity and potential risks, may already be outpacing humanity's current capacity for understanding and control.
Business dilemmas and technological fears may seem distant, but they share a common root: we are conducting a cognitive revolution with industrial-era thinking. We desire AI to be as cheap and accessible as tap water (hence the push for free access), while simultaneously hoping it will be as powerful and magical as a spell (hence the relentless scaling). In reality, it currently resembles an extremely energy-intensive, precision instrument requiring careful maintenance. The user exodus triggered by Doubao's pricing is the market teaching us to recognize AI's "cost"; while Anthropic's warning is science reminding us that the true "cost" may extend far beyond the figures on a bill.
Even more ironically, while the industry remains immersed in grand narratives of "building models" and "building applications," AI has quietly breached the most hardcore academic sanctum. It solved a mathematical problem that had baffled mathematicians for 80 years. Beneath this news, headlines like "Mathematicians are panicking" carry a mocking tone but precisely pierce a deeper anxiety. If AI can conquer the forefront of human intellect, then where lie the boundaries of the rationality, creativity, and even the very thought we take pride in? This is no longer a parable of tools surpassing humans, but a preview of "cognitive substitution" in action. It compels us to rethink: in a world where intelligence can independently discover theorems, where should humanity's unique value anchor be placed?
From the plummeting commercial data of monthly active users, to the philosophical warning of "self-improvement," to the cognitive shock of solving a mathematical problem, a fantastical yet realistic portrait of AI development emerges. We are hastily scrambling to find business models for it while fearing its potential disruption; we are awed by its capabilities while anxious about our own place. The entire industry is like a young person suddenly endowed with immense power but unsure how to wield it, spinning in a vortex of commerce, ethics, and technology. The introduction of charges is just the beginning—the ultimate interrogation of cost, boundaries, and responsibility has only just begun to unfold. And the answer may lie in the departure or retention of the next 6.1 million users, or in the birth of some groundbreaking code.
Disclaimer: The above content is generated by AI and is for reference only.