前段时间，雷科技揭晓了年度"灯塔产品"，那些代表行业顶尖水准的设备赢得了掌声与认可。然而，灯塔之下也会有阴影，科技圈的历史从来都是由成功与 试错共同书写的。所以，在盘点完辉煌之后，雷科技编辑部经过多轮激烈的内部讨论与投票，最终筛选出了这份"年度失望榜单"。 这些产品或许并不算差，但是却与外界对它的期待有明显差距，或是与前代相比显得过于平庸甚至退步。这个榜单并非为了否定这款产品，而是为了督促大 家更多地关注产品本身的失败原因，并从中吸取教训做出更好的产品。 毕竟，看清脚下的坑，未来的路才能走得更稳，正如那句老话所言："失败乃是成功之母"。 Sora 2：含着金钥匙出生，奈何用力太猛 而且，随着Sora 2进一步缩紧免费以及Plus付费会员的额度，下调视频生成效果并频繁改动用户权益，很多用户都开始转向别的AI视频平台，根据第三方应 用市场监测平台SensorTower的统计数据，Sora 2的30天用户留存率不到1%，60天留存率接近0。 虽然这并不能完全代表Sora 2的用户留存率，但也足以说明OpenAI在Sora 2的发布、运营等方面出现了重大失误。另外，随着Gemini对ChatGPT发出挑战， Open ...

Core Insights - A new wave of AI-native companies is emerging globally, focusing on AI as a core product or service from inception, differentiating them from companies that merely integrate AI into existing operations [2] - The research identifies three main development routes for AI consumer hardware: AI-native exploration, gradual enhancement of existing devices, and model-centric empowerment [3][4][5] - The AI consumer hardware market is witnessing significant innovation, with new product categories like AI phones, smart glasses, and companion robots rapidly gaining traction [3] Group 1: AI Consumer Hardware Development Routes - Route 1: AI-native devices, exemplified by products like Rabbit R1 and Humane AI Pin, aim to redefine interaction paradigms but face challenges in user experience and market acceptance [3][4] - Route 2: Companies like Apple and Meta represent a gradual enhancement approach, integrating AI capabilities into existing devices to improve user experience while maintaining brand strength [4][11] - Route 3: Model-centric companies like OpenAI focus on empowering various devices through APIs and SDKs, allowing for widespread integration of AI capabilities without building proprietary hardware [5][6] Group 2: Emerging Business Models - The AI-native exploration model relies on high-margin hardware and subscription services, targeting niche markets but struggling with user adoption due to functionality limitations [8][9] - The gradual enhancement model emphasizes hardware sales and value-added subscription services, benefiting from established brand recognition and user familiarity [11][12] - The model empowerment approach mirrors the Android ecosystem, focusing on API-based monetization and enterprise-level services, but faces challenges in cost and integration [13][14] Group 3: Trends and Future Outlook - The integration of AI models with hardware is becoming increasingly important, with companies collaborating with chip manufacturers to optimize performance across devices [15][16] - The trend towards "unobtrusive" interaction is evident, with AI glasses and other devices aiming to enhance user experience without replacing smartphones [17][18] - Long-term, the ultimate form of AI consumer hardware may evolve into a more integrated and seamless user experience, with AI acting as a primary interface for various applications [21][22]

Core Viewpoint - The article discusses the emergence of AI-native companies that prioritize artificial intelligence as their core product or service, leading to new technologies, products, and business models in the AI hardware industry [2]. Group 1: AI Consumer Hardware Development Routes - AI consumer hardware has seen significant innovation in 2023, with new categories like AI phones, smart glasses, rings, headphones, and companion robots rapidly emerging [4]. - The development routes can be categorized into three main paths: 1. AI-native devices exploring new interaction paradigms, represented by products like Rabbit R1 and Humane AI Pin, which rely on semantic understanding and task execution driven by large models [5]. 2. Gradual enhancement of existing devices with AI capabilities, exemplified by Apple and Meta, which integrate AI into established hardware like smartphones and wearables [6]. 3. Model-centric empowerment paths led by companies like OpenAI, focusing on providing AI capabilities through APIs and SDKs to third-party devices [7]. Group 2: Emerging Business Models in AI Consumer Hardware - The article identifies the initial emergence of business models corresponding to the three development routes, highlighting their respective core challenges: 1. AI-native exploration models rely on high-priced hardware and subscription services to generate stable revenue streams, but face challenges in proving hardware value and user adoption [10]. 2. Gradual enhancement models focus on hardware sales and value-added subscription services, benefiting from low user recognition barriers and high market acceptance [12]. 3. Model empowerment paths replicate aspects of the Android model, charging for API access and enterprise-level services, but face challenges in cost and adaptation to various hardware [15]. Group 3: Future Trends in AI Consumer Hardware - The integration of upstream and downstream in the industry is becoming tighter, with model vendors collaborating with chip manufacturers to optimize model performance across devices [18]. - The trend towards "unobtrusive" interaction is accelerating hardware paradigm shifts, with AI glasses becoming a focal point for competition among tech giants and emerging brands [21]. - Long-term, AI hardware is expected to evolve towards a model where AI acts as a primary interface, with voice and natural language interactions becoming the norm, potentially replacing traditional graphical user interfaces [27].