8 Point 1氪丨Analyst Exposes Apple Vision Pro Product Line Being Removed; Jensen Huang to Appear on Variety Show for the First Time; Fenbi CEO Apologizes for Calling Unemployed College Students 'Deserve to Not Find Work' 8点1氪丨分析师曝苹果Vision Pro产品线被移除；黄仁勋将首次亮相综艺节目；粉笔CEO就骂人大学生“活该找不到工作”道歉

While Silicon Valley debates whether glasses should have screens, Jensen Huang has flown to Seoul, raising glasses with Korean chaebols at a barbecue joint and even planning to appear on Yoo Jae-suk's variety show. The NVIDIA CEO's "entertainment debut" is no spur-of-the-moment decision. It's a natural extension of the intensifying battle for AI compute: collaborations with SK and LG target the AI semiconductor supply chain, while variety show exposure directly battles for the public's mindshare. A tech leader's personal charisma is becoming a more effective business weapon than technical specs. Bill Gates appeared on this same show last year; now it's Huang's turn. This has become a ritual: if you want to sell chips in the Asian market, you first need to learn how to dance with the entertainment industry. This effortless cross-industry grace itself announces a new rule: in the AI era, a CEO must simultaneously be a product manager, diplomat, and celebrity.

But the world of AI deployment isn't all glamorous. Chalk CEO Zhang Xiaolong's speech at the School of Philosophy at Renmin University of China turned into a classic case study of poor entrepreneurial EQ. Flaunting wealth, insulting students by saying they "deserved to be jobless"—these remarks stand in absurd contrast to his identity as CEO of an education technology company. Education should empower, yet he used the most brutal means to humiliate. What's even more ironic is that this disaster unfolded at a top-tier university's philosophy school—a place meant to explore human value and dignity. No matter how sincere the apology letter, it cannot erase the harsh words caught on tape. This incident brutally reveals that while many tech nouveau riche are sprinting ahead on capital, their humanistic literacy and empathy remain in a primitive stage. Technical capability does not equate to the ability to be a good human—a lesson every tech company needs to learn.

Meanwhile, WeChat's collaboration with manufacturers like Huawei and Xiaomi to launch Agent-to-Agent assistant capabilities might be the most noteworthy undercurrent on this list. When you can directly tell your phone's voice assistant, "Use WeChat to video call Mom," it means a breach has been blown in the wall between the mobile operating system and the super app. This is no longer a simple API call—it's direct dialogue between intelligent agents. Honor devices can already perform this operation via YOYO. Behind this lies WeChat's reimagining of traffic entry points: it doesn't want to just be an app you open; it wants to become the ubiquitous "service air." For phone manufacturers, this is a golden opportunity to boost user stickiness; for WeChat, it's a moat defending against erosion by super apps like Douyin. The most subtle aspect of this collaboration is that it breaks China's long-standing "walled garden" rules—and the catalyst is AI. When AI becomes the interaction hub, closed ecosystems must find open interfaces.

On another front, ByteDance's four goals for 2026—world models, video generation, coding, and Doubao commercialization—clearly outline the company's anxieties and ambitions. The world model aims to rival Google's Genie 3, video generation must maintain its lead, coding seeks to build a closed development loop, and commercialization targets the office scenario. The level of detail in this roadmap reads almost like a military plan. Of particular note is the phrase "Coding dogfooding": letting AI itself refine programming tools, using the company's own products to train itself. If this "eat your own dog food" strategy succeeds, it could create a powerful data flywheel. ByteDance is transforming from a content distribution giant into an infrastructure company for the AI era, and commercialization in the office scene is clearly the most handy lever for penetrating the B2B market.

While tech giants map out the future in boardrooms, a UN report delivers a stark warning: by 2030, AI data centers will consume as much water as 1.3 billion people need for a year. As we marvel at AI-generated HD videos and smooth conversations, we're invisibly overdrafting the planet's blue blood. Data from China Literature shows that AI short dramas have become the largest token-consuming scenario, accounting for 55%—more than double the consumption of coding. This is pure dark humor: humanity is using massive energy and water resources primarily to produce melodramatic three-minute episodes. The value direction of technology is being ruthlessly twisted by the market—what drives instant traffic and ad revenue is often the most superficial application. Goldman Sachs predicts SpaceX will "burn through" $350 billion by 2030, with 80% going to AI. At this scale of capital investment, if it ultimately serves only entertainment-to-death content and convenience for a select elite, we may need to recalculate the true cost of this technological revolution.

Returning to WeChat's A2A collaboration: it may symbolize a more pragmatic path for AI implementation—not chasing flashy sci-fi vibes, but solving fundamental needs like "how to more easily video call Mom." While giants race to build world models and pursue the grand narrative of AGI, the real revolution may be happening in these capillary-level connections. When different AI assistants can shake hands, converse, and relay information on our behalf, a more complex and convenient digital society will truly begin to take shape. Just don't forget: behind every intelligent interaction lies real costs in energy consumption, data privacy, and humanistic ethics. As the wheels of technology roll forward, we must both admire the new landscapes they reveal and stay vigilant about the tracks they leave behind.

苹果终于承认了一个尴尬的事实：人们根本不想在脑袋上戴个昂贵又笨重的头盔去“进入未来”。最新供应链调查显示，Vision Pro产品线已被从路线图中彻底抹去，资源转向预计2027年上市的无显示屏AI眼镜。这个决定迟到了整整两年。那个号称“空间计算元年”的雄心壮志，如今看来更像一场耗资巨大的行为艺术。市场用钱包投了票：两万块人民币的价格，换来的是在飞机上被邻座侧目的尴尬，和在办公室里被同事当作异类的目光。苹果的转向不是创新乏力，而是终于看清了现实——在当下，人们想要的不是“另一个屏幕”，而是真正能融入生活、不那么引人注目的智能辅助。从这个角度看，2027年那副“无屏”眼镜，或许才是苹果对“可穿戴设备”这个概念的第一次认真作答。

当硅谷在纠结眼镜该不该有屏幕时，黄仁勋已经飞往首尔，在烤肉店与韩国财阀把酒言欢，还准备登上刘在石的综艺节目。黄教主的“娱乐圈首秀”绝非心血来潮。这背后是AI算力争夺战白热化的必然延伸：与SK、LG的合作直指AI半导体供应链，而综艺曝光则是对公众心智的直接争夺。科技领袖的个人魅力，正在成为比技术参数更有效的商业武器。比尔·盖茨去年上过这个节目，现在轮到黄仁勋，这已经形成一种仪式——想在亚洲市场卖芯片，先得学会和娱乐工业共舞。这种跨界游刃有余的姿态，本身就在宣告一种新规则：在AI时代，CEO必须同时是产品经理、外交官和明星。

但AI落地的世界并不光鲜。粉笔CEO张小龙在中国人大哲学院的演讲翻车，堪称一堂经典的企业家情商反面教材。炫富、辱骂学生“活该找不到工作”，这些言行与他教育科技公司CEO的身份形成荒诞对比。教育本该是赋能，他却用最粗暴的方式进行羞辱。更讽刺的是，这场灾难发生在顶级学府的哲学院——这里本应探讨人的价值与尊严。道歉信写得再诚恳，也抹不掉录音里那些刺耳的字句。这事件赤裸裸地揭示：许多科技新贵在资本上狂飙突进，在人文素养和共情能力上却还停留在原始阶段。技术能力不等于做人能力，这是所有科技公司都该补上的一课。

而微信与华为、小米等厂商合作推出Agent-to-Agent助手能力，可能是这份列表里最值得关注的暗线。当你可以直接对着手机语音助手说“用微信给妈妈打视频电话”，这意味着手机操作系统与超级应用之间的墙被炸开了一个缺口。这不再是简单的API调用，而是智能体之间的直接对话。荣耀机型已经能用YOYO完成这个操作。这背后是微信对流量入口的重新想象：它不想只做一个被打开的App，而是想成为无处不在的“服务空气”。对手机厂商而言，这是增加用户粘性的绝佳机会；对微信而言，这是防御抖音等超级应用侵蚀的护城河。这场合作最微妙之处在于，它打破了中国互联网长期以来“围墙花园”的规则，而催化剂正是AI——当AI成为交互中枢，封闭的生态就必须找到开放的接口。

另一边，字节跳动为2026年立下的四个目标——世界模型、视频生成、编程、豆包商业化——清晰地勾勒出这家公司的焦虑与野心。世界模型要对标Google Genie 3，视频要继续领跑，编程要打通开发闭环，商业化则瞄准了办公场景。这份路线图的详尽程度，近乎一份军事计划书。尤其值得注意的是“Coding的Dogfooding”这个表述：让AI自己来打磨编程工具，用自家产品训练自己。这种“吃自己的狗粮”策略若能成功，将形成强大的数据飞轮。字节正在从内容分发巨头，转型为AI时代的基础设施公司，而办公场景的商业化，显然是它撬动B端市场最顺手的支点。

当科技巨头们在会议室里谋划未来时，联合国一份报告却给出了冰冷的警告：到2030年，AI数据中心的耗水量将相当于13亿人一年的基本用水。我们一边惊叹于AI生成的高清视频和流畅对话，一边却在无形中透支着地球的蓝色血液。中文在线的数据显示，AI短剧已成为最大的Token消耗场景，占比55%，是编程消耗的两倍以上。这简直是一个黑色幽默：人类用海量的能源和水资源，首要用途竟然是生产那些三分钟一集的狗血短剧。技术的价值方向正在被市场无情地扭曲——最能带来即时流量和广告收入的，往往也是最肤浅的应用。高盛预测SpaceX到2030年将“烧掉”3500亿美元，其中80%用于AI。这种规模的资本投入，如果最终只服务于娱乐至死的内容和少数精英的便利，我们恐怕需要重新计算这场技术革命的真正成本。

回到微信那个A2A合作。它或许象征着一种更务实的AI落地路径：不追求炫酷的科幻感，而是解决“如何更方便地给妈妈打电话”这种最基本的需求。在巨头们争相构建世界模型、追求AGI宏大叙事的同时，真正的变革可能正发生在这些毛细血管级的互联中。当不同的AI助手能替我们握手、交谈、传递信息时，一个更复杂也更便利的数字社会才真正开始成形。只是别忘了，在每一次智能交互的背后，都有真实的能源消耗、数据隐私和人文伦理代价。技术的车轮向前碾压，我们既要欣赏它开辟的新风景，也要时刻警惕它留下的碾痕。