Core Insights - Memories.ai, founded by Shawn Shen, has launched LVMM 2.0, a Large Visual Memory Model, and announced a partnership with Qualcomm for native operation on Qualcomm processors by 2026 [2][9] - The company focuses on developing AI's visual memory capabilities, having completed a $8 million seed round in July 2025, led by Susa Ventures with participation from notable investors like Samsung Next and Fusion Fund [2] Company Background - Founder Shawn Shen holds a PhD in Engineering from Trinity College, Cambridge, and previously worked as a core research scientist at Meta Reality Labs, focusing on human-computer interaction and augmented reality [3] - Co-founder Ben Zhou also has experience at Meta Reality Labs, working on AI assistants for Meta's Ray-Ban glasses [3] - Eddy Wu has been appointed as the Chief AI Officer, bringing five years of experience from Meta, where he was involved in GenAI research [3] Product Development - LVMM 2.0 was released three months after the first generation, maintaining performance while reducing parameter count by 90%, making it more suitable for edge devices [6] - The model converts raw video into structured memory on-device, addressing video searchability issues by encoding and compressing frames to create an index for millisecond retrieval [7][8] Technical Advantages - Running on Qualcomm processors significantly reduces latency, lowers cloud costs, and enhances data localization for improved security [8] - The model integrates video, audio, and images to provide contextual results, ensuring a consistent experience across devices like smartphones and cameras [8] Applications - Practical applications of LVMM 2.0 include enhancing AI capabilities in smart glasses, security systems, and robots, enabling real-time understanding and response [11]

机器之心报道 机器之心编辑部 Dr. Shawn Shen ，联合创始人兼首席执行官（左）； Ben (Enmin) Zhou ，联合创始人兼首席技术官（右）。 由 前 Meta Reality Labs 顶尖科学家团队创立的 AI 研究实验室 Memories.ai ，正式宣布 完成 8 00 万美元 种子轮融资 。本轮融资由 Susa Ventures 领投，三星风投 （Samsung Next）、Fusion Fund 等知名机构跟投。 Memories.ai 团队已经在大模型领域完成了一项重大的突破成果，剑指 AI 系统的「记忆缺失」问题，为视觉模型创造了强大的「记忆大脑」。 「最 强大脑 」 众所周知， 大模型是标标准准的「金鱼记忆」 。 比如，大多数 AI 系统都缺乏对历史画面的记忆，难以理解前后之间的关联。 就像我们经常开的玩笑，「记忆是个先进先出栈」，只不过大模型的栈容量似乎总是不够用。 这种「金鱼记忆」限制了它们在需要深入理解场景和动态变化的应用中发挥作用，尤其是在视频密集型任务里表现不佳。 为了彻底解决这个问题， Memories.ai 通过其核心创新 —— 大视觉记忆模型（LVMM），为 ...