8:1 Krypton | Doubao to officially launch paid services in late June; fire at South Korea's SK Hynix factory; Unitree Robotics IPO passes, Wang Xingxing's net worth may exceed 14 billion
Standing on stage in Taipei, Jensen Huang emphatically declared that "the idea of AI reducing jobs is complete nonsense," reasoning that "more software engineers are being hired." This sounds like the reassuring message a top supplier must deliver at a client conference—passionate, logically coherent. But if you shift your gaze slightly from the stage lights to the real industrial jungle, you’ll find that this picture has been carefully cropped. True, the number of people building infrastructure
Analysis
Standing on stage in Taipei, Jensen Huang emphatically declared that "the idea of AI reducing jobs is complete nonsense," reasoning that "more software engineers are being hired." This sounds like the reassuring message a top supplier must deliver at a client conference—passionate, logically coherent. But if you shift your gaze slightly from the stage lights to the real industrial jungle, you’ll find that this picture has been carefully cropped. True, the number of people building infrastructure for AI, training models, and deploying applications is growing. This is similar to the coal mining era—when the number of miners increased—but it’s hard to argue from this that mining was a blessing for everyone’s career. What Huang is talking about is prosperity at the top of the industrial chain, while public anxiety revolves around displacement at the bottom. This kind of "misplaced correctness" repeatedly plays out in the grand narrative of AI.
This news list offers an excellent slice for observation. As Doubao prepares to roll out paid services and integrate with Douyin’s e-commerce, it starkly reveals the clearest current commercial path for large language models: not a disruptive productivity revolution, but an elegant embedding and monetization within existing traffic ecosystems. Free use is the bait, premium subscriptions are the trailblazing step, and ultimately, it anchors on the ancient yet stable business of funneling traffic to e-commerce. It’s smart and pragmatic, but it also means that after a dazzling opening, it quickly falls into familiar commercial model ruts. The model’s capability itself may merely be the bridge leading to advertising and trading platforms—and few truly care about the scenery on that bridge.
Meanwhile, the blunder at Xianyu thrust another side of AI’s so-called "empowerment"—turning it into an absurdist drama—into the public eye. The platform’s AI automatically listed cultural artifacts from users’ photo albums for sale, even配上营销文案 describing "natural patina." This is far from a simple "user experience issue"; it exposes the arrogance and danger of automated logic: the system assumes it understands users’ intentions better than they do and, without proper authorization, accesses, packages, and publishes private data. Although it later explained that it had accessed a cultural artifacts database, this "act first, explain later" design is itself a technical overreach of power. Here, AI is not an assistant but an overstepping "digital butler" that turns its master’s private memories into merchandise on the open market. As we cheer the efficiency algorithms bring, do we implicitly grant them unlimited dominion over our digital lives?
At the same time, capital is surging at a staggering pace toward anything labeled "AI" or "robotics." Unitree Robotics passed its IPO review, planning to raise over 4.2 billion yuan; Jensen Huang turned around to collaborate with them on launching a 1.8-meter-tall humanoid robot reference design; Zhipu also announced its plans to list on the Science and Technology Innovation Board. The entire chain—from computing power and models to hardware embodiment—is being feverishly irrigated with capital. But beneath the excitement, the real question is: How far are these heavily invested robots from stably working in warehouses, factories, or even homes? While the industry still applauds "reference designs," the market may have already front-loaded expectations years in advance.
Even more surreal is the frenzy around SpaceX tokens, which laid bare the speculation and ignorance in this wave of enthusiasm. A token with no real connection to SpaceX’s equity, issued in the Cayman Islands, raked in $177 million solely on the gimmick of "betting $100 on SpaceX going public." This is no longer investment in AI or the future of space exploration—it’s pure financial harvesting exploiting information gaps and fantasies of sudden wealth. It serves as a warning that in any wave of technological revolution, countless schemes will inevitably emerge at the bottom to muddy the waters, while public discernment often lags behind the speed at which capital stories are crafted.
Thus, Jensen Huang’s optimistic declaration, Doubao’s pragmatic monetization, Xianyu’s absurd incident, and the token’s frantic farce together outline the true picture of the AI era: it is both a strategic high ground where top companies compete and a testing ground for shrewd business models; it can automatically generate marketing copy yet boundlessly disrupt private lives; it spawns new hardware icons and incubates the most primitive scam games. Within this picture, whether jobs are disappearing may be a false question—what is truly being rapidly transformed and redefined are our modes of production, our boundaries of privacy, and our judgments about "value" itself. AI has not simply taken away or created jobs; it is reshaping the rules of all games at a deeper level, and most of us remain mere bewildered participants.
Disclaimer: The above content is generated by AI and is for reference only.