OpenAI在做一个会自己动的智能音箱，不是普通音箱，用GPT-Live实时对话还能识别物体和人。和苹果打官司抢人，这瓜可以吃一吃。
OpenAI首款硬件设备被描述为无屏幕智能音箱，配备可自主移动的机械部件。设备使用摄像头和传感器理解环境，基于GPT-Live实现同时听和说功能。它能学习用户习惯并利用个人数据（如邮件）主动行动，售价区间200-300美元，最早2027年发货。苹果于7月10日提起商业秘密诉讼，指控OpenAI挖角其工程师加快硬件开发。
OpenAI's first hardware device is described as a s…
OpenAI's first hardware device is described as a screenless smart speaker with mechanical parts that move on their own. This is the most interesting bit. First time I read the news I missed this fact. But it's what sets it apart from another static voice box.
Latest reporting says it will be battery powered so you can carry it between rooms, uses cameras and sensors to understand context, runs on GPT-Live for simultaneous listening and speaking, and learns from habits plus personal data like emails to act proactively. This comes after the 2025 acquisition of Jony Ive's io startup, with LoveFrom handling design. Earlier reporting from February pointed to a built-in camera for recognising objects and people, a rough price range of $200 to $300, and shipping no earlier than 2027.
At first glance this overlaps heavily with existing home speakers. Most owners will probably just plug it in and leave it in one spot rather than moving it around. A few units scattered through the house would make more practical sense than one portable unit. The mechanical movement stands out as the real attempt to give the AI physical presence and personality.
Apple filed its trade secrets lawsuit on 10th July, just four days before these details surfaced, alleging a pattern of poaching ex-employees and information transfer aimed at speeding up OpenAI's hardware plans. The timing feels convenient given how much the project draws on former Apple hardware talent, though the underlying tensions predate the latest leak and Apple had already raised private concerns back in February. OpenAI maintains the device differs significantly from Apple's own products. Even so, the legal process itself can create delays and uncertainty around manufacturing and timelines regardless of how the claims ultimately play out.
I'm disappointed it won't have a display. My Nest Max device is well used precisely because of its display. It's great for feedback so you know what is heard correctly and what is being said out loud, plus all the visual confirmations for actions. A display would also be perfect for generating images, showing videos, avatar and all the usual things you'd want from a screen.
That said, having ChatGPT in a hardware device is very exciting. I'm not concerned about the portability. That would only matter temporarily while we can only get one or two devices. Once stock is available, I'd probably get three or four and dot them around my house anyway.
There are more devices coming from OpenAI. Something, I hope, that is more portable than a box for your home. But by next year, you would really expect Gemini for Home to be way better by then and be more competitive. https://t.co/8kCPDXeIj5