AI News AI资讯 9h ago Updated 2h ago 更新于 2小时前 49

Apple will let you build workflows using AI in its new Shortcuts app 苹果将在新的快捷指令应用中使用AI构建工作流程

Apple’s revelation at WWDC 2026 that it’s injecting AI into Shortcuts to make it "easier" isn’t just a feature update; it’s a philosophical admission that the power-user tool it proudly built for a decade was, by design, a failure for most of its intended audience. The core of the problem was always one of translation: the gulf between human intent ("let me know when I’m heading home") and the rigid, block-by-block logic of visual scripting. By letting you describe what you want in plain English 当苹果在WWDC 2026的舞台上,由一位高管略带歉意地承认自家“快捷指令”用起来“确实感觉……挺复杂的”时,台下不知有多少老用户会心一笑。这迟来的坦诚,比任何新功能的介绍都更真实。而苹果给出的解药,依旧是他们如今最时髦的万灵药:人工智能。具体来说,是让Apple Intelligence去理解一句人话,然后替你把那些让人头晕的变量、操作和条件判断给拼凑出来。

75
Hot 热度
70
Quality 质量
65
Impact 影响力

Analysis 深度分析

Apple’s revelation at WWDC 2026 that it’s injecting AI into Shortcuts to make it "easier" isn’t just a feature update; it’s a philosophical admission that the power-user tool it proudly built for a decade was, by design, a failure for most of its intended audience. The core of the problem was always one of translation: the gulf between human intent ("let me know when I’m heading home") and the rigid, block-by-block logic of visual scripting. By letting you describe what you want in plain English, Apple is finally trying to build a bridge across that gulf, powered by Apple Intelligence. But in doing so, it’s trading one set of frustrations for a potentially more insidious one: the illusion of simplicity masking a loss of control and understanding.

Let’s be clear about the victory here. The old Shortcuts was a sandbox for tinkerers, a delight for the subset of people who think in terms of variables, conditional statements, and API calls. For everyone else, it was a wall of intimidating blocks with names like "Get Contents of URL" and "Set Variable." Apple’s marketing VP admitting the process "can feel, well, complicated" is the tech equivalent of a car company admitting its dashboard has too many buttons. It’s a refreshingly honest starting point. The promise of typing "Send my wife a text with my ETA when I leave the office" and having a functional shortcut materialize is genuinely powerful. It lowers the barrier to automation from "learn a new skill" to "state your desire." This could finally move automation from a niche hobby to a mainstream utility, weaving it into the fabric of daily digital life in a way it never has before.

However, this is where my skepticism spikes. This move is quintessentially Apple: take a complex, powerful tool, sand down all the sharp edges, put a friendly AI face on it, and then present the streamlined version as the whole story. The danger lies in what happens when the AI misinterprets your intent. In the block-based Shortcuts, an error was a logical puzzle you could debug. You could see exactly where the flow broke. With an AI-generated shortcut, what happens when it mishears "leaving the office" as "leaving any location"? Does it send a text every time you step out for coffee? Do you then have to learn the block-based editor anyway to fix the AI’s mistake, but now you’re debugging an opaque, AI-generated workflow you never built? The tool risks becoming a black box that generates other black boxes. The learning curve isn’t eliminated; it’s just relocated to a higher level of abstraction where troubleshooting is more mystifying.

Furthermore, this feels like a critical step in Apple’s long-term strategy to make its operating system less a platform for apps and more a single, intelligently curated environment. By routing automation through Apple Intelligence, it reinforces the centrality of its own AI. Shortcuts that pull data from third-party apps or use non-Apple services might become second-class citizens in this new paradigm. The example given—using Apple Maps and Messages—is perfectly seamless. But what about automating a workflow that involves Slack, Notion, and Google Calendar? Will the AI seamlessly understand those, or will it gently nudge you toward using Apple’s equivalents? This isn’t just about convenience; it’s about ecosystem control. Making the "easy path" the one that runs deepest within Apple’s walled garden is a business play as much as a user-experience one.

There’s also a question of empowerment versus dependency. The old Shortcuts, for all its complexity, taught you something. You learned how your phone’s systems talked to each other. You gained a tangible skill. The new AI-driven version risks creating a generation of users who are dependent on a "magic" layer they don’t understand. It’s the difference between learning to cook a meal from a recipe versus pressing a button on a meal-replacement machine. The outcome might be similar, but the knowledge and adaptability are worlds apart. When the AI inevitably fails to understand a nuanced request, that user will have no recourse but to abandon the idea, having never been given the tools to build it themselves.

Ultimately, this update is a pragmatic capitulation. Apple saw that the most powerful automation tool in its OS was being used by a sliver of its customers. The choice was to either accept that reality or try to democratize it. They’ve chosen the latter, which is commendable. But it’s a solution that trades transparent complexity for opaque simplicity. The real test won’t be whether the AI can build the basic "ETA to wife" shortcut. It will be whether it can handle the messy, specific, and often illogical requests of real human life, and whether it can do so while still giving the user a sense of agency and understanding. If it can, it’s a revolution. If it can’t, it’s just a shinier, more frustrating layer of digital bureaucracy.

当苹果在WWDC 2026的舞台上,由一位高管略带歉意地承认自家“快捷指令”用起来“确实感觉……挺复杂的”时,台下不知有多少老用户会心一笑。这迟来的坦诚,比任何新功能的介绍都更真实。而苹果给出的解药,依旧是他们如今最时髦的万灵药:人工智能。具体来说,是让Apple Intelligence去理解一句人话,然后替你把那些让人头晕的变量、操作和条件判断给拼凑出来。

这本质上是一次对“复杂性”的精巧转嫁。快捷指令从诞生之日起,就背负着一个尴尬的定位:它是为“高级用户”设计的平民自动化工具。这个定义本身就是个悖论。它试图降低自动化的门槛,却因为提供了一套功能强大但逻辑晦涩的图形化编程界面,而筑起了新的高墙。过去,你想实现“下班离开工司自动发消息给伴侣”,需要理解位置触发、获取地址、调用地图、计算时间、最后通过信息App发送这一整套链路。现在,你只需要像使唤一个听话但可能有点笨的助理一样说:“嘿,我下班了就告诉我对象我大概几点到家。” 系统会帮你搞定剩下的。

这听起来很美好,是苹果典型的“魔法时刻”叙事。但细究之下,魔法的背后是更深层的控制与定义权移交。当你用自然语言描述需求时,你在依赖AI对你意图的“理解”,而这种理解被苹果的算法严格框定。AI需要从你模糊的表述中,决定调用哪些API,遵循什么逻辑顺序。用户从“流程的构建者”退化成了“需求的提出者”。极客们乐在其中的,正是那种亲手搭建、掌控每一步的“构建感”。现在,苹果说:别费劲了,我来替你想。这对大众市场是福音,但对原本那批核心用户而言,或许是一种“降维打击”——他们被自己熟悉的工具“优化”出局了。

更值得玩味的是苹果在AI策略上的精明与局限。他们没有像竞争对手那样,急于推出一个无所不知的通用聊天机器人与Siri竞争。而是选择让AI如盐入水,溶解进系统最基础的应用层里。在邮件里智能总结、在写作时帮你润色、现在又在快捷指令里充当“翻译官”。这是一种极其务实且风险较低的路径:AI不直接成为主角,而是作为提升现有功能体验的“增强剂”。这很苹果,永远追求软硬件生态体验的丝滑与整合。

但这也暴露了其根本性的谨慎甚至保守。苹果的AI,目前看更像一个能力超强、知识渊博但被严格限制在自家院子里干活的管家。它能帮你把杂事处理得井井有条,但你别指望它能跳出这个院子,带来认知上的颠覆或创造力上的迸发。它解决的是“如何做”的效率问题,而非“做什么”或“为什么做”的创意与决策问题。快捷指令的AI化,让非技术用户也能触达自动化,这扩大了其用户基本盘。但与此同时,工具的灵魂可能也被稀释了。一个完全靠自然语言生成的快捷指令,它的每一步逻辑对你而言是黑箱,调试和修改的难度可能比之前更高。你得到了“易用性”,却失去了“可理解性”与“完全控制力”。

那个下班自动通知的例子很暖心,也很实用,但它恰恰是这类“描述性自动化”的甜区:步骤明确、数据源固定、流程线性。一旦需求变得复杂、需要嵌套多层逻辑或处理异常情况,AI是否还能如此可靠地“翻译”用户的心思?它会不会生成一段冗长、低效甚至包含冗余步骤的“正确但笨拙”的指令?届时,真正的极客用户恐怕还是会撸起袖子,回到那个熟悉的、虽然复杂但绝对可靠的手动编辑界面。

所以,苹果用AI改造快捷指令,是一次明智的市场拓宽,也是一次对工具原始精神的微妙背叛。它让自动化从一项“技能”变成了一种“服务”。这或许会是未来所有复杂工具的终极归宿:藏起所有的齿轮与扳手,只给你一个输入框和输出结果。我们得到了更大的便利,也默认了更多的黑箱。对于99%的普通用户而言,这无疑是进步。而对于剩下那1%的人来说,他们可能开始怀念那个需要自己动手连接每一个“动作”和“变量”的、有点“笨”但无比清晰的时代了。苹果再次证明了它定义大众科技产品的能力,但在这个过程中,某些属于极客的乐趣与掌控感,也正在被优雅地格式化掉。

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。

产品发布 产品发布 编程 编程 对话系统 对话系统
Share: 分享到: