正值大模型从"算力堆叠"走向"推理优先"的关键节点，清华大学智能产业研究院（AIR）创始院长、中国工程院外籍院士张亚勤提出： 新一轮人工智能，是信息智能、物理智能和生物智能的融合，本质上也是原子、分子和比特的融合。 也就是说，在规模定律持续发挥作用的前提下，当参数规模、数据体量与算力资源跨过某个阈值，智能就不再只停留在模式识别，而是开始"涌现"—— 先是从鉴别式 AI 走向生成式 AI，再从生成式 AI 走向以智能体为代表的新范式。 在本次量子位 MEET2026 智能未来大会上，他也将 ChatGPT 和 DeepSeek，视作这一轮演进中的两个重要里程碑： 从 ChatGPT 到 DeepSeek，AI 正沿着"智能 +"的路径进入新一轮浪潮。 至于未来 5~10 年的主战场，在他看来，将走向"智能体互联网"时代 —— 基础大模型像操作系统一样在全球范围内收敛到不超过 10 个；而智能体会取代 今天的大部分 SaaS 和 App，成为企业和个人与世界交互的默认形态，同时这也是通往 AGI 的必经之路。 为了完整体现张亚勤的思考，在不改变原意的基础上，量子位对演讲内容进行了编辑整理，希望能给你带来更多启发。 ...

从ChatGPT到DeepSeek，AI正沿着"智能+"的路径进入新一轮浪潮。 正值大模型从"算力堆叠"走向"推理优先"的关键节点，清华大学智能产业研究院（AIR）创始院长、中国工程院外籍院士张亚勤提出： 新一轮人工智能，是信息智能、物理智能和生物智能的融合，本质上也是原子、分子和比特的融合。 也就是说，在规模定律持续发挥作用的前提下，当参数规模、数据体量与算力资源跨过某个阈值，智能就不再只停留在模式识别，而是开始"涌现"—— 先是从鉴别式AI走向生成式AI，再从生成式AI走向以智能体为代表的新范式。 在本次量子位MEET2026智能未来大会上，他也将ChatGPT和DeepSeek，视作这一轮演进中的两个重要里程碑： 前者通过统一表征与token化，把文本、语音、图像乃至蛋白质、点云等数据纳入同一空间； 后者则以高效率、高性能、低价格和开源路径，把大模型从"预训练时代"推向以推理为核心的"DeepSeek时刻"。 基础大模型像操作系统一样在全球范围内收敛到不超过10个；而智能体会取代今天的大部分SaaS和App，成为企业和个人与世界交互的默认形态，同时这 也是通往AGI的必经之路。 为了完整体现张亚勤的思考 ...

正值大模型从"算力堆叠"走向"推理优先"的关键节点， 清华大学智能产业研究院（AIR）创始院长、中国工程院外籍院士张亚勤 提出： 编辑部 整理自 MEET2026 量子位 | 公众号 QbitAI 从ChatGPT到DeepSeek，AI正沿着 "智能+" 的路径进入新一轮浪潮。 新一轮人工智能，是信息智能、物理智能和生物智能的融合，本质上也是原子、分子和比特的融合。 也就是说，在规模定律持续发挥作用的前提下，当参数规模、数据体量与算力资源跨过某个阈值，智能就不再只停留在模式识别，而是开 始"涌现"—— 先是从鉴别式AI走向生成式AI，再从生成式AI走向以 智能体 为代表的新范式。 在本次 量子位MEET2026智能未来大会 上，他也将ChatGPT和DeepSeek，视作这一轮演进中的两个重要里程碑： 前者通过统一表征与token化，把文本、语音、图像乃至蛋白质、点云等数据纳入同一空间； 后者则以高效率、高性能、低价格和开源路径，把大模型从"预训练时代"推向以推理为核心的"DeepSeek时刻"。 至于未来5~10年的主战场，在他看来，将走向 "智能体互联网"时代 —— 基础大模型像操作系统一样在全球范围内 ...

Group 1 - Danone is restructuring its leadership in China, appointing Zhou Zhigang to lead its specialized nutrition business starting January 1, 2026, as part of its "Revitalizing Danone" strategy [1][2] - The restructuring aims to create a more streamlined leadership structure to accelerate transformation and enhance focus and agility in operations [1] - Danone's CEO has identified medical nutrition as a significant long-term growth opportunity, with adult medical nutrition in China experiencing double-digit growth annually for over a decade [2] Group 2 - Nestlé is also merging its Wyeth Nutrition and Nestlé infant nutrition businesses into a single unit, effective January 1, 2026, with Joel Seah as the head [3] - The infant formula market is under pressure, with Nielsen IQ reporting a growth rate of only 0.6% in the second quarter of 2025, and a decline in marriage registrations in China by over 20% in 2024 [4] - Despite foreign brands maintaining a brand advantage, domestic brands like Feihe and Yili have surpassed foreign companies in market share as of 2022 [4][5]

Core Insights - The key point of the news is that the CTO of Huolala, Zhang Hao, emphasized the importance of applying AI capabilities to business scenarios rather than just building foundational large models. Huolala was recognized for its innovative AI applications in logistics at the "WISE2025 Business King" conference [1][2]. Group 1: AI Application and Impact - Huolala has developed an AI safety control system that monitors every step from order placement to transportation completion, achieving a 30% reduction in daily risk orders for hazardous goods transportation and a 100% identification rate for risk orders [2]. - The usage rate of Huolala's AI Coding exceeds 90%, with over 60% penetration of AI in the R&D process, significantly enhancing research and development efficiency [2]. - The "Photo Goods Selection" feature allows users to take a picture of their goods, enabling AI to recommend suitable vehicle types with a maximum error margin of less than 10% and an average error of less than 10 centimeters [2]. Group 2: Future Directions and Strategy - Zhang Hao stated that the rapid iteration of foundational large models means that many current issues may not be problems in the near future. The company aims to leverage end-to-end large model assistants for tasks like intelligent vehicle selection and internal operations [3]. - The focus for Huolala will continue to be on deepening application scenarios and building platform capabilities to convert technology into business value [3].

Core Viewpoint - The company, Yunzhisheng (09678), has made significant progress in the commercialization of its foundational large model and intelligent agent platform, collaborating with well-known domestic enterprises to implement projects valued at nearly RMB 200 million, with delivery planned by the end of 2025 [1] Group 1 - The company has established deep cooperation with domestic partners to advance its foundational large model and intelligent agent projects [1] - The implementation of these projects is based on the company's proprietary computing scheduling platform, Shanhai large model, and Shouya Agent intelligent platform [1] - This progress signifies a solid step forward in the commercialization process of the company's foundational large model and intelligent agent platform [1] Group 2 - The advancements further validate the company's technological leadership in foundational large models and intelligent agent platforms [1] - The company plans to deepen its research and commercialization efforts in the smart living sector, leveraging its strong advantages [1]

Core Insights - The company, Yunzhisheng (09678.HK), has made significant progress in the commercialization of its foundational large model and intelligent agent platform [1] - The company has partnered with well-known domestic enterprises to implement projects valued at nearly RMB 200 million, utilizing its proprietary computing scheduling platform, Shanhai large model, and Shouya Agent intelligent platform [1] - The delivery of these projects is planned to be completed by the end of 2025, indicating a solid step forward in the commercialization process [1] Company Developments - The advancements signify the recognition of the company's technological leadership in foundational large models and intelligent agent platforms [1] - The company aims to further deepen its research and commercialization efforts in the smart living sector, leveraging its strong advantages [1]

Core Insights - The company, Yunzhisheng (09678), has made significant progress in the commercialization of its foundational large model and intelligent agent platform [1] - The company has partnered with well-known domestic enterprises to implement projects valued at nearly RMB 200 million, with delivery planned by the end of 2025 [1] - This advancement indicates the company's leading position in the technology of foundational large models and intelligent agent platforms [1] Summary by Categories Commercialization Progress - The company has achieved substantial progress in the commercialization of its foundational large model and intelligent agent platform [1] - The implementation of projects worth approximately RMB 200 million is underway, showcasing the company's ability to attract significant partnerships [1] Technological Leadership - The advancements signify further recognition of the company's technological leadership in foundational large models and intelligent agent platforms [1] - The company plans to deepen its research and commercialization efforts in the smart living sector, leveraging its strong advantages [1]

UNISOUND AI TECHNOLOGY CO., LTD. 雲知聲智能科技股份有限公司 （於中華人民共和國註冊成立的股份有限公司） （股份代號：9678） 香港交易及結算所有限公司及香港聯合交易所有限公司對本公告的內容概不負責，對其準確性或完整性亦不發表 任何聲明，並明確表示概不就因本公告全部或任何部分內容而產生或因倚賴該等內容而引致的任何損失承擔任何 責任。 本公告乃由雲知聲智能科技股份有限公司（「本公司」）自願作出，旨在知會本公司股東 及潛在投資者有關本公司的最新業務發展情況。 本公司董事會（「董事會」）欣然宣佈，本公司基礎大模型和智能體平台商業化取得實質 性進展，近日與國內知名企業等合作夥伴深入合作，將基於自有知識產權的算力調度平 台、山海大模型以及獸牙Agent智能體平台，實現價值近人民幣2億元的基礎大模型和智 能體項目實施部署，計劃將於2025年底前完成交付。 該進展意味著本公司在基礎大模型和智能體平台商業化落地進程中邁出了堅實的一步， 同時也標誌著公司基礎大模型和智能體平台技術的領先性得到進一步認可。本公司將 在具備深厚優勢的智慧生活領域進一步深耕大模型和智能體的研發及商業化落地。 承董事會 ...

Core Insights - The rapid development of the AI industry is accelerating iterations across various sectors, presenting significant industrial opportunities [3] - The scale of the AI industry is projected to be at least 100 times larger than the previous generation, indicating substantial growth potential [5] Group 1: Trends in AI Development - The first major trend is the transition from discriminative AI to generative AI, now evolving towards agent-based AI, with task lengths doubling and accuracy exceeding 50% in the past seven months [7] - The second trend indicates a slowdown in the scaling law during the pre-training phase, with more focus shifting to post-training stages like reasoning and agent applications, while reasoning costs have decreased by 10 times [7] - The third trend highlights the rapid advancement of physical and biological intelligence, particularly in the intelligent driving sector, with expectations for 10% of vehicles to have L4 capabilities by 2030 [7] Group 2: AI Risks and Industry Structure - The emergence of agent-based AI has significantly increased AI risks, necessitating greater attention from global enterprises and governments [8] - The fifth trend reveals a new industrial structure characterized by foundational large models, vertical models, and edge models, with expectations for 8-10 foundational large models globally by 2026, including 3-4 from China and the same from the U.S. [8] - The future is anticipated to favor open-source models, with a projected ratio of 4:1 between open-source and closed-source models [8]