OpenAI 获得新“技能”：ChatGPT 从 Claude 的剧本中学习

OpenAI 的 ChatGPT 将升级“技能”，反映 Claude 量身定制的 AI 功能，向用户承诺在 AI 领域提供更个性化和更强大的交互。

The Big Apple's tech scene, always buzzing, has its eyes firmly on the latest development from the AI heavyweights. It seems OpenAI, the company that brought ChatGPT to the masses, is taking a leaf right out of rival Anthropic's playbook. Word on the street, or rather, on the X, is that ChatGPT is getting its own version of "Skills," a feature that's already made Claude quite the savvy assistant.

纽约的科技界总是热闹非凡，它们密切关注着人工智能巨头的最新发展。 OpenAI，这家将 ChatGPT 带给大众的公司，似乎正在效仿竞争对手 Anthropic 的做法。街上的消息，或者更确切地说，X 上的消息是，ChatGPT 正在推出自己版本的“技能”，这一功能已经使 Claude 成为了相当精明的助手。

ChatGPT's New Bag of Tricks

ChatGPT 的新技巧

For those of us who've been tinkering with AI, we know custom GPTs have been handy. But imagine something even more integrated, more intuitive. That's where "Skills" come in. Claude's "Skills" are essentially super-charged, folder-based instructions that teach the AI specific abilities and workflows. Think of it: a frontend design skill plugin for Claude, making it a virtuoso at understanding your web app's interface while you're busy coding up a storm. Pretty slick, right?

对于我们这些一直在研究 AI 的人来说，我们知道自定义 GPT 非常方便。但想象一下更集成、更直观的东西。这就是“技能”的用武之地。克劳德的“技能”本质上是超级强大的、基于文件夹的指令，教授人工智能特定的能力和工作流程。想一想：Claude 的前端设计技能插件，使其成为在您忙于编码时理解 Web 应用程序界面的大师。很光滑，对吧？

Now, ChatGPT is reportedly joining the club. Codenamed "hazelnuts" – because, you know, everything's better with a secret snack name – these new "Skills" are slated to arrive as slash commands. We're talking a dedicated Skills editor and even the ability to convert your existing custom GPTs into these potent new capabilities. While the exact rollout date is still a bit of a mystery, whispers suggest we might see them in action around January 2026. Plenty of time to mentally prepare for an even smarter chatbot.

据报道，现在 ChatGPT 也加入了该俱乐部。代号为“榛子”——因为，你知道，有了秘密零食名称，一切都会变得更好——这些新的“技能”预计将以斜线命令的形式出现。我们谈论的是专用的技能编辑器，甚至是将现有的自定义 GPT 转换为这些强大的新功能的能力。虽然确切的推出日期仍然有点神秘，但有传言称我们可能会在 2026 年 1 月左右看到它们投入使用。有足够的时间为更智能的聊天机器人做好心理准备。

The Method Behind the Madness

疯狂背后的方法

Why the sudden move to mimic a competitor? Well, in the cutthroat world of AI, standing still is a recipe for disaster. OpenAI has been raking in the dough, sure – with compute margins hitting a respectable 70% and the ChatGPT mobile app pulling in billions. But even with those impressive numbers, and despite widespread adoption that sees hundreds of millions of users interacting with ChatGPT weekly, there's always the relentless drumbeat of competition. Sam Altman himself reportedly emphasized a "critical situation" internally, urging accelerated improvements as Google and Anthropic nip at their heels.

为什么突然采取模仿竞争对手的行动？好吧，在人工智能的残酷世界中，原地踏步只会酿成灾难。当然，OpenAI 一直在赚大钱——计算利润达到了可观的 70%，ChatGPT 移动应用也带来了数十亿美元的收入。但即使有这些令人印象深刻的数字，尽管ChatGPT 得到了广泛采用，每周有数亿用户与ChatGPT 进行交互，但竞争始终存在。据报道，萨姆·奥尔特曼本人在内部强调了“危急情况”，敦促加速改进，因为谷歌和 Anthropic 紧随其后。

This push for "Skills" isn't just about keeping up; it's about deepening the utility. As AI becomes more integrated into our daily lives – from shopping to planning to even health research, as one report suggested – the demand for specialized, context-aware assistance skyrockets. Generic answers won't cut it when you need an AI that truly understands your niche workflow or domain.

对“技能”的推动不仅仅是跟上；而是要跟上。这是为了深化实用性。正如一份报告所指出的，随着人工智能越来越融入我们的日常生活——从购物到规划，甚至健康研究——对专业的、情境感知的帮助的需求猛增。当您需要一个真正了解您的利基工作流程或领域的人工智能时，通用答案并不能解决问题。

My take? This is a savvy strategic pivot. By adopting a "Skills" framework, OpenAI is acknowledging that the future of AI isn't just about raw power, but about precision and personalization. It’s about turning a generalist into a specialist, making AI not just a helpful tool, but an indispensable expert tailored to your exact needs. The evidence is clear: when an AI can analyze something as complex as crypto market dynamics (as ChatGPT itself did for Ethereum vs. Solana), you know there's untapped potential for specialized applications. Giving it explicit "Skills" just supercharges that capability.

我的看法是？这是一个精明的战略支点。通过采用“技能”框架，OpenAI 承认人工智能的未来不仅仅在于原始能力，还在于精度和个性化。它将通才变成专家，使人工智能不仅是一个有用的工具，而且是根据您的具体需求量身定制的不可或缺的专家。证据很明确：当人工智能能够分析像加密货币市场动态这样复杂的东西时（就像 ChatGPT 本身对以太坊与 Solana 所做的那样），你就知道专业应用程序有尚未开发的潜力。赋予它明确的“技能”只会增强这种能力。

What's Next for Our Digital Pals?

我们的数字朋友的下一步是什么？

So, there you have it. ChatGPT is getting "Skills," and it sounds like our digital companions are about to get a whole lot more discerning. Whether you're coding up a storm or just need an AI that truly 'gets' your workflow, these new capabilities promise to make interactions smoother and smarter. It's a bit like giving your trusty old car a souped-up engine and a custom paint job. The future's looking bright, and increasingly skilled, for our AI pals. Just don't ask them to fetch you a pastrami on rye; they're not quite there yet. But hey, a New Yorker can dream!

所以，你就知道了。 ChatGPT 正在获得“技能”，听起来我们的数字伴侣即将变得更加敏锐。无论您是在编码，还是只需要一个真正“掌握”您的工作流程的人工智能，这些新功能都有望使交互变得更加流畅和智能。这有点像给您值得信赖的旧车配备增强型发动机和定制油漆工作。对于我们的人工智能朋友来说，未来看起来是光明的，而且技术也越来越熟练。只是不要让他们给你拿一份黑麦熏牛肉；他们还没有完全做到这一点。但是，嘿，纽约人可以梦想！