UBS Halves Asia ESG Workforce in Global Restructuring 瑞银在全球重组中将亚洲ESG员工人数减半

Remember the original OpenAI that championed "benefiting all of humanity" and was initially non-profit? It was the embodiment of our utopian imagination for AI. Today, the blueprint for this "super app" rests on a cold, clear core logic: user stickiness, usage duration, and paid conversions. The gradually filling icon bar beside the chat window doesn't reflect the dawn of Artificial General Intelligence (AGI), but rather the success playbook of TikTok, WeChat, or other Super Apps. OpenAI is stepping down hard from its pedestal as a "thinking partner," squeezing into the crowded race to become a "digital life manager." This, in itself, is understandable. The irony, however, is that it has shattered its own initially crafted, simplest yet most powerful product philosophy: focusing on one interaction and perfecting it.

Behind this abrupt turn is the physical manifestation of growth anxiety. As the novelty fades and daily active user growth curves begin to plateau, investor pressure from the likes of Microsoft for returns penetrates through the code, striking at the product's core. This is less about building a "one-stop solution" for users and more about constructing a "one-stop harvesting scheme" for its own revenue growth curve. It's desperate to prove it isn't just a useful but disposable tool, but a platform that can infinitely expand and accommodate all needs. This obsession with "platformization" is a familiar story in tech history, but it's often the beginning of a product becoming bloated and losing its soul. Today we cheer its integration of PPT creation and data analysis, tomorrow we may have to endure the forced embedding of e-commerce recommendations and short-video feeds.

A contrasting and interesting parallel is the news of UBS cutting half its ESG (Environmental, Social, and Governance) staff in Asia. This seemingly unrelated piece of information tears open another slice of reality: in the face of macroeconomic uncertainty and downward pressure, all the "politically correct" narratives of long-termism, value investing, and beautiful stories can be re-evaluated before the next quarterly report. The axe UBS wielded against its ESG department sends a stark and cruel signal—when survival becomes the primary concern, departments that are "correct" but offer no short-term returns are often the first to be cut. The frantic charge in the AI sector and the contraction in "correctness" from traditional financial giants together paint a ukiyo-e of capital swaying between idealism and reality.

Then there's the story of the Korean exchange halting buy orders due to program trading. It reads like a parable, reminding us that no matter how grand the narrative of the upper-layer application, the underlying system always follows a different set of physics. Algorithm-driven trading can trigger circuit breakers, and similarly, a product transformation driven by algorithms and commercial objectives can also trigger a "circuit breaker" in user trust. GPT's overhaul is a redeployment of algorithms and products, but once the user's mindset forms the perception that "it's changed, it's become utilitarian," this intangible risk of a "trading halt" might be harder to mend than any technical flaw.

So, what exactly has the GPT overhaul changed? On the surface, it's added a row of draggable icons and integrated DALL·E 3 and a code interpreter. At its core, it's OpenAI's complete metamorphosis from a "research institution" into a "tech company." The AI partner, which occasionally refused to answer questions and was fraught with idealistic contradictions, is being trained into a smoother, more commercialized, and frankly more boring digital servant. We might gain more convenient tools, but we are likely losing forever a dialogue partner who once filled us with wonder and deep thought.

The headline "Chat is Dead" might be overly sensational, but the pure, simple chat interface that initially stunned the world with ChatGPT—its minimalist soul—is indeed dissipating. On the congested highway to becoming a "super app," OpenAI has discarded its sharpest weapon—focus and restraint. It has chosen a broader, but perhaps more mediocre, path. And for us users, this may mean learning anew how to interact with AI within a "super app" meticulously designed to extract more of our time, data, and attention. Is this progress, or just another classic cycle that begins with ideals and ends with traffic?

当OpenAI把“史上最大改版”与“从单一聊天变身超级应用”这些重磅词砸向市场时，我们听到的不是技术革命的号角，更像是一家备受压力的公司，在财报和董事会注视下，被迫转向时发出的齿轮摩擦声。ChatGPT的这次“重生”，与其说是愿景的进化，不如说是商业模式的紧急迫降。

还记得那个以“造福全人类”为旗号，最初甚至非盈利的OpenAI吗？它曾是我们对AI乌托邦想象的具象。而今，这个“超级应用”的蓝图，核心逻辑冰冷而清晰：用户粘性、使用时长、付费转化。聊天框旁边那个逐渐填满的图标栏，映照出的不是通用人工智能（AGI）的曙光，而是TikTok、微信或Super App的成功学模板。OpenAI正奋力从“思考伙伴”的神坛走下，挤进“数字生活管家”的拥挤赛道。这本身无可厚非，但讽刺的是，它亲手打碎了自己最初构建的那个最简洁、最强大的产品哲学：专注于一个交互，做到极致。

这场急转的背后，是增长焦虑的实体化。当新鲜感褪去，日活用户增长曲线开始平缓，微软等投资者要求回报的压力便穿透代码，直达产品核心。与其说这是在为用户构建“一站式解决方案”，不如说是在为自身的收入增长曲线构建“一站式收割方案”。它急切地想证明，自己不只是一个好用但用完即走的工具，而是一个可以无限扩张、容纳一切需求的平台。这种对“平台化”的执念，在科技史上屡见不鲜，但往往也是产品走向臃肿、失去灵魂的开端。今天我们为它整合PPT制作和数据分析而欢呼，明天就可能要忍受它强行嵌入的电商推荐和短视频信息流。

与此形成有趣对照的，是瑞银在亚洲将ESG（环境、社会和治理）员工砍掉一半的消息。这则看似无关的资讯，恰好撕开了另一个现实切口：在宏观的不确定性与经济下行压力面前，所有关于长期主义、价值投资和美好叙事的“政治正确”，都可能在下一个季度财报前被重新评估。瑞银的裁员刀挥向ESG部门，传递的信号直白而残酷——当生存成为首要议题，那些“正确”但短期难见回报的部门，往往首当其冲。AI领域的狂飙突进，与传统金融巨头在“正确性”上的收缩，共同绘制了一幅资本在理想与现实间摇摆的浮世绘。

再看那条韩国交易所因程序交易暂停买盘的新闻。它像一则微型寓言，提醒我们无论上层应用的故事讲得多么天花乱坠，底层系统永远遵循着另一套物理规则。算法驱动的交易可以引发熔断，同样，由算法和商业目标驱动的产品转型，也可能引发用户信任的“熔断”。GPT的改版是一次算法和产品的重新部署，但用户的心智一旦形成“它变了，变功利了”的认知，这种无形的“停牌”风险，或许比任何技术漏洞都更难修复。

所以，GPT的改版到底改了什么？表面上，是增加了一排可以拖拽的图标，集成了DALL·E 3和代码解释器。骨子里，是OpenAI从一家“研究机构”向“科技公司”的彻底蜕变。那个偶尔会拒绝回答、充满理想主义矛盾感的AI伙伴，正在被训练成一个更顺滑、更商业化、也更无趣的数字服务员。我们或许会得到更便捷的工具，但可能正在永久失去一个曾让我们感到惊奇与深思的对话者。

“Chat已死”这个标题或许过于煽情，但那个最初以纯粹聊天界面惊艳世界的ChatGPT，其简洁的产品灵魂确实正在消散。在通往“超级应用”的拥堵高速路上，OpenAI扔掉了自己最锋利的武器——专注与克制。它选择了一条更宽广，也可能更平庸的道路。而对于我们这些用户，这或许意味着要在一个被精心设计来榨取我们更多时间、数据和注意力的“超级应用”里，重新学习如何与AI相处。这究竟是进步，还是一场始于理想，终于流量的经典轮回？