最新一期的 The VergeCast 节目由 David Pierce 主持，嘉宾 David Amell 深入探讨了正在兴起的 AI 驱动设备趋势，这些设备声称“不是手机”，但却与手机惊人地相似。讨论的核心在于，据报道由 OpenAI、SpaceX 和亚马逊等公司开发的这些设备，能否真正取代智能手机，抑或它们注定只会变成智能手机本身。
Pierce 指出了一个悖论：那些旨在将个人科技从智能手机中解放出来的公司，最终却常常造出与手机非常相似的东西。他认为，其核心动机是将 AI 置于用户生活的中心位置，并绕过苹果和谷歌已形成的双头垄断。Amell 同意“AI 优先设备”这个 *想法* 在概念上是说得通的，但实际实现起来却仍难以捉摸。
主持人探讨了所提出的“范式转变”——从以应用程序为中心、“基于工具”的手机使用方式转向“基于意图”的交互，即自然语言 AI 处理跨各种服务的请求。尽管承认当前在不同应用间切换的低效率以及 Apple 快捷指令等解决方案的局限性，Pierce 质疑默认使用聊天窗口的界面是否真的具有吸引力。Amell 指出，对于多步任务，物理按钮当前比冗长的 AI 提示更高效。
讨论中的一个关键时刻是 Amell 描述了他个人使用 Siri 管理演唱会门票购买的经历，其中涉及网络搜索、提醒和跨平台操作。这个真实世界的例子展示了深度集成 AI 的潜力，证明了当这些功能无缝运行时，用户行为能多么迅速地改变。然而，Pierce 很快指出，苹果绝不会允许第三方 AI 实现如此深度的集成，这迫使其他公司不得不构建自己的硬件和生态系统。
对话触及了“智能手机陷阱”，即像裴宇（Carl Pei）的 Nothing 这样的公司意识到，它们必须先“造一部手机”才能建立生态系统。这引发了一个问题：据传“基于 Android”的 OpenAI 设备，是否可能只是一个配备了 AI 重点启动器（launcher）的普通安卓手机，类似于问题重重的 Rabbit R1。
亚马逊据传的新手机，代号“Transformer”，受到了特别批评。Pierce 引用报道称，它将优先集成亚马逊服务（Prime Video、Amazon.com、Grubhub），这让他将其贴上了“资本主义操作系统”的标签。两位主持人都认为亚马逊没有从 Fire Phone 的失败中吸取任何教训，没有人想要这种消费主义驱动的设备。
Pierce 提出了他唯一可行的理论：或许这些“非手机”只是“计算和网络连接的砖块”，作为通过耳塞或智能眼镜等配件进行 AI 交互的管道。Amell 反驳说，如果真是这样，公司就应该直接制造这些配件，因为它们成本更低，也更容易被接受。
最终，主持人对这三家公司（SpaceX、OpenAI、亚马逊）在 AI 核心智能手机方面的成功可能性进行了排名：
1. **SpaceX AI：** Amell 将 SpaceX 排在第一位，理由是其负面品牌印象被稀释（火箭很酷，盖过了其他争议）以及埃隆·马斯克（Elon Musk）已有的粉丝群。Pierce 同意马斯克有将项目“硬是推进”成功的历史。
2. **OpenAI：** 排在第二位，与 2023 年最初的炒作相比，其潜力有所减弱。用户热情已经消退，许多高级用户已将 AI 集成到他们现有的设备中。OpenAI 还面临着关于 AI 社会影响的重大品牌挑战。
3. **亚马逊：** 排在最后，其“资本主义操作系统”的方法被认为是一场“灾难”，表明其根本不了解用户需求。
讨论以一个悬而未决的问题结束：如果不是应用程序，也不是聊天窗口，那么与 AI 交互的下一个 UI 范式会是什么，尤其是在移动设备上？
The latest episode of The VergeCast, hosted by David Pierce with guest David Amell, delves into the burgeoning trend of AI-centric devices that claim *not* to be phones, yet bear striking resemblances to them. The discussion centers on whether these devices, reportedly being developed by companies like OpenAI, SpaceX, and Amazon, can genuinely unseat the smartphone or if they are destined to simply become smartphones themselves.
Pierce highlights the paradox: companies aiming to revolutionize personal tech away from the smartphone often end up making something very phone-like. The core motivation, he argues, is the desire to put AI at the forefront of users' lives, bypassing the established duopoly of Apple and Google. Amell agrees that the *idea* of an AI-first device makes conceptual sense, but the practical implementation remains elusive.
The hosts explore the proposed "paradigm shift" from app-centric, "tool-based" phone usage to "intent-based" interaction, where natural language AI handles requests across various services. While acknowledging the inefficiency of current app-hopping and the limitations of solutions like Apple Shortcuts, Pierce questions whether an interface that defaults to a chat window would be truly compelling. Amell notes the current efficiency of physical buttons for multi-step tasks versus verbose AI prompts.
A key moment in the discussion comes when Amell describes his personal experience using Siri to manage concert ticket purchases, which involved web searches, reminders, and cross-platform actions. This real-world example showcases the potential of deeply integrated AI, demonstrating how quickly user behavior can change when such capabilities work seamlessly. However, Pierce quickly points out that Apple would never allow a third-party AI to achieve such deep integration, forcing other companies to build their own hardware and ecosystems.
The conversation touches on the "smartphone trap," where companies, like Carl Pei's Nothing, realize they "have to make a phone" first to establish an ecosystem. This leads to the question of whether the rumored OpenAI device, reportedly "Android-based," might just be a regular Android phone with an AI-focused launcher, similar to the problematic Rabbit R1.
Amazon's rumored new phone, codenamed "Transformer," draws particular criticism. Pierce reads reports suggesting it would prioritize integration with Amazon services (Prime Video, Amazon.com, Grubhub), leading him to label it "Capitalism OS." Both hosts conclude that Amazon has learned nothing from the failure of the Fire Phone and that nobody wants such a consumerism-driven device.
Pierce offers his sole viable theory: perhaps these "not-phones" are simply "bricks of compute and network connectivity," acting as conduits for AI interactions occurring through accessories like earbuds or smart glasses. Amell counters that if that's the case, companies should just build the accessories directly, as they are lower cost and easier to adopt.
Ultimately, the hosts rank the three companies (SpaceX, OpenAI, Amazon) on their likelihood of success with an AI-centric smartphone:
1. **SpaceX AI:** Amell places SpaceX first, citing the dilution of negative brand perception (rockets are cool, overriding other controversies) and the established fanbase of Elon Musk. Pierce agrees that Musk has a history of "horsing" projects into existence.
2. **OpenAI:** Ranked second, their potential is seen as diminished compared to the initial hype of 2023. User enthusiasm has waned, and many power users have already integrated AI into their existing devices. OpenAI also faces significant branding challenges regarding AI's societal impact.
3. **Amazon:** Ranked last, the "Capitalism OS" approach is deemed a "disaster," showing a fundamental misunderstanding of what users want.
The discussion concludes with the lingering question: if not apps and not a chat window, what *is* the next UI paradigm for interacting with AI, especially on a mobile device?