AI News AI资讯 15h ago Updated 1h ago 更新于 1小时前 53

OpenAI says 'chat is dead' and plans to rebuild ChatGPT as a full-blown agent app OpenAI称'聊天已死',计划将ChatGPT重建为完整的代理应用

“Chat is dead.” So declare the gleeful pallbearers at OpenAI, preparing to bury the very thing that made them famous. They’re not just building a better chatbot; they’re attempting a full organ transplant, hoping to replace its conversational heart with the robotic gizmos of an “agent.” The news that ChatGPT is being reimagined as a “superapp” – a Frankenstein bundle of coding tools, autonomous agents, and embedded partner apps like Canva and Booking.com – is less a product update and more a phi “聊天已死”,OpenAI内部的这句话,像是一声提前敲响的丧钟,宣告了ChatGPT赖以成名的模式即将被抛弃。但仔细一品,这口号本身,就充满了一种硅谷式的话术狡诈和战略模糊。说“聊天已死”,他们却要做的,是把ChatGPT变成一个包罗万象的“超级应用”——而实现这一切最核心、最底层的交互界面,**依然是那个该死的聊天框**。你让编程代理帮你写代码,让设计代理调用Canva,让旅行代理预订酒店,靠的什么?不还是用自然语言去“聊天”下达指令吗?所以,“死”的不是聊天这个交互形式,而是我们最初对“聊天”的浪漫想象:一个可以进行思想漫游、开放性探索的对话伙伴。OpenAI正亲手将这个伴侣,打造成一个雷

80
Hot 热度
70
Quality 质量
75
Impact 影响力

Analysis 深度分析

“Chat is dead.” So declare the gleeful pallbearers at OpenAI, preparing to bury the very thing that made them famous. They’re not just building a better chatbot; they’re attempting a full organ transplant, hoping to replace its conversational heart with the robotic gizmos of an “agent.” The news that ChatGPT is being reimagined as a “superapp” – a Frankenstein bundle of coding tools, autonomous agents, and embedded partner apps like Canva and Booking.com – is less a product update and more a philosophical admission of exhaustion. The pure, elegant, text-in/text-out paradigm that captivated the world is, apparently, a dead end. What a bizarre and telling epitaph.

The pivot reeks of a classic tech-industry syndrome: the terror of being a utility. After the initial firework display, the question for any killer app becomes, “How do we become the platform?” Just as Facebook feared being a mere photo-sharing site and WeChat grew into an ecosystem for ordering taxis and paying bills, OpenAI now sees the chat interface as a constraining garden wall. The goal is no longer to answer your question brilliantly, but to become the transactional layer that does the thing you’re asking about. It’s the difference between a brilliant librarian who explains how to find a flight and a travel agent who just books it for you.

But here’s my sharp take: in fleeing from the specter of being a “mere chatbot,” OpenAI risks alienating the very thing that gave it power—the illusion of a thinking partner. ChatGPT’s magic wasn’t in its ability to call an API; it was in the fluency, the reasoning, the seeming cognition. An “agent” that silently spins in the background, clicking through Booking.com interfaces to fulfill a vague command, trades that magic for mechanical utility. It’s a demotion. We’re no longer having a conversation with a seemingly intelligent entity; we’re barking orders at a glorified macro. The personality gets optimized away in favor of the task flow. It’s a sad trade.

And let’s talk about this “superapp” ambition. This isn’t visionary; it’s a cargo cult imitation of Asian tech giants like WeChat or Grab, where everything from payments to food delivery lives inside one chat interface. But Western digital behavior is fundamentally different. We are conditioned to use specialized apps. The “superapp” strategy only works when you have near-monopolistic control over a user’s digital identity and payment layer, something OpenAI is nowhere near achieving. Attempting to graft a Canva editor or a Booking.com widget directly into the ChatGPT window is a recipe for UI clutter, not convenience. It’s the “portal” fallacy of the early 2000s web, reborn with a Transformer backbone.

The most cynical reading of this move is that it’s a revenue play dressed in futurism. A pure chatbot is hard to monetize beyond subscriptions. But an “agent platform” that takes a cut of every booking, every code deployment, every design collaboration facilitated through its system? That’s a marketplace. That’s a tollbooth. OpenAI is building the mall and hoping to charge rent to every store inside it. The “chat is dead” mantra isn’t an observation; it’s a justification for why they need to become a middleman in every digital transaction.

The risk they’re ignoring is trust. For an agent to act autonomously—handling my finances, booking my travel, writing and deploying code—it requires a level of trust that borders on the absurd. Who is liable when an AI agent books you on a flight that doesn’t exist, or deploys code that takes your company’s server down? The current model of chat-and-confirm is a built-in, human-in-the-loop safety mechanism. Removing it in favor of autonomous action isn’t just a technical challenge; it’s a legal and ethical minefield. We’ve seen enough AI hallucinations to know that giving the keys to the kingdom to a probabilistic model is insanity. The agent future they’re selling requires a 99.999% reliability that doesn’t exist.

Furthermore, this smells of desperation to stay ahead. Google, Anthropic, and a dozen startups are breathing down their necks. The raw performance gap between leading models is closing. So, what’s left? Features. Bundling. Ecosystem. It’s the Microsoft playbook: if you can’t win on elegance, win on integration. They’re hoping to create so much sticky, embedded utility that users can’t leave, even if a competitor has a slightly smarter brain. It’s a defensive maneuver masquerading as an offensive one.

At the end of the day, the “chat is dead” proclamation feels premature, almost naive. The conversational interface is the most natural human-computer interface ever devised. Its potential is far from exhausted. What OpenAI seems to be discarding is the difficult part—building robust, safe, and delightful conversation—in favor of the flashy part—demoing an agent that automatically sends emails. They’re chasing the shiny object of automation while undermining the core interaction that made people care in the first place.

They’re not killing chat. They’re abandoning it for the chance to become the operating system of your digital life. It’s a monumental bet that users don’t want a brilliant conversationalist, but an invisible, task-crushing servant. I, for one, will miss the conversation. I suspect I’m not alone. The most interesting question isn’t whether agents will work, but what gets lost in the translation when the chatbot stops talking and starts doing.

“聊天已死”,OpenAI内部的这句话,像是一声提前敲响的丧钟,宣告了ChatGPT赖以成名的模式即将被抛弃。但仔细一品,这口号本身,就充满了一种硅谷式的话术狡诈和战略模糊。说“聊天已死”,他们却要做的,是把ChatGPT变成一个包罗万象的“超级应用”——而实现这一切最核心、最底层的交互界面,依然是那个该死的聊天框。你让编程代理帮你写代码,让设计代理调用Canva,让旅行代理预订酒店,靠的什么?不还是用自然语言去“聊天”下达指令吗?所以,“死”的不是聊天这个交互形式,而是我们最初对“聊天”的浪漫想象:一个可以进行思想漫游、开放性探索的对话伙伴。OpenAI正亲手将这个伴侣,打造成一个雷厉风行、但略显刻板的“任务管家”。

这步棋,野心昭然若揭。他们不想只做一个惊艳世界的“功能”,他们想做一个不可或缺的“平台”,一个AI时代的操作系统。捆绑Coding工具、Canva、Booking.com,这套组合拳打出去,瞄准的是用户每天的数字生活轨迹。写邮件、做图、安排旅行、处理工作流……如果这些动作都能在一个ChatGPT的“壳”里闭环完成,那用户的留存时长和依赖度将呈指数级增长。这才是“超级应用”的算盘:用便捷性取代专业性,用一站式服务筑高围墙,最终将海量用户的注意力和数据,牢牢锁在自己的生态里。从产品角度看,这是微软“全家桶”策略在AI时代的激进变体;从商业角度看,这是将万亿参数模型的惊人能力,迅速变现、建立护城河最直接的路径。

然而,光鲜的愿景下,藏着开发者生态的隐忧与用户自由的让渡。当ChatGPT的“代理人”们开始直接调用Canva的设计API或Booking.com的预订服务时,那些曾经在插件商店里努力获客的独立开发者们,其价值和入口被严重稀释了。OpenAI等于是在说:各位,你们不需要自己造轮子了,来我的高速公路上跑吧,但我就是这条公路的唯一收费站和交通规则制定者。用户看似获得了便利,实则交出了选择权。你未来处理创意任务,可能不再有十种不同的AI设计工具可选,而只有ChatGPT调用的那一种“推荐方案”。这种便捷的代价,是数字生活可能性的收窄和创新的同质化。

更深一层看,“聊天已死,代理当立”的转向,暴露了OpenAI在商业模式焦虑下的路径依赖。纯聊天机器人难以规模化收费,也难以量化其带来的具体价值。但“代理”则不同,它可以清晰地与“生产力提升”挂钩,可以按任务完成度、流程优化程度来定价,可以更容易地包装成企业解决方案卖出高价。这本质上,是将一个充满可能性的通用智能,强行塞进一个名为“效率工具”的紧身衣里。它或许能解决很多实际问题,但它也主动关闭了通往更奇异、更发散、更人性化的AI应用可能性的大门。我们最终得到的,可能是一个无比强大但略显无趣的“数字工头”。

所以,“聊天已死”?不,它只是被夺舍了。它依然活着,但灵魂已从一个充满好奇的对话者,变成了一个目标明确的执行者。OpenAI的这次豪赌,赌的是用户对“便利”的绝对臣服,赌的是平台经济在AI领域的终极胜利。但作为用户,我们或许该偶尔停下来问自己:当每一次打开AI,都是为了“完成某件事”而不是“探索某种可能”时,我们是否在无形中,也为自己画上了一个名为“效率”的牢笼?

Disclaimer: The above content is generated by AI and is for reference only. 免责声明:以上内容由 AI 生成,仅供参考。

GPT GPT Agent Agent 产品发布 产品发布 大模型 大模型 代码生成 代码生成
Share: 分享到: