作 者丨肖潇 实习记者隆欣玲 编 辑丨王俊 美国宣布对所有贸易伙伴加征"对等关税"的消息持续动荡，这几天里，手机里的新闻 弹窗爆炸，不同地区、不同行业的关键词在标题里轮番滚动。 想 要 快 速 看 懂 发 生 了 什 么 ， 却 越 刷 越 眼 花 缭 乱 ， 突 然 想 到 ： 能 不 能 让 A I 帮 忙 总 结 一 下"美国最新关税加征政策对市场的影响"？ A I果然高效。短短几秒里，它就给出了股市情况、行业冲击、中国应对措施。有言之 凿凿的数据，有生动的案例，比如华为启动了"鸿蒙供应链计划"，联合1 5 0 0家供应商 构 建 去 美 化 产 业 链 ；Temu 、 S H EI N 等 平 台 被 迫 提 价 1 5 % ~ 2 5 %；Ti kTo k 商 家 伪 装 东 南亚店铺销售. . . . . . 但这些"故事"这么快就出炉了吗？作为记者，出于职业本能的半信半疑，一条条点进 链接查看，结果发现有的说法出自个人账号，看不出来源；有的是好几年前的行业数 据——今年情况早就不同了；还有的数据根本就是无中生有，前文提到的几则信息均 是如此。 这并非偶然。就像一滴墨染入清水，A I编造的内容正在 ...

Core Viewpoint - The article discusses the launch of Genspark, a new AI agent developed by MainFunc, which has gained attention for its capabilities in automating tasks and outperforming competitors in benchmark tests [4][8]. Group 1: Genspark Overview - Genspark is a general-purpose AI agent that can autonomously think, plan, and utilize tools to assist users with daily tasks [11]. - It emphasizes real-time interaction, reduced hallucination and error rates, and allows users to make adjustments during the process [12]. - The agent can perform complex tasks, such as converting a 5-hour YouTube video into a 10-page PPT [13]. Group 2: Company Background - MainFunc, the company behind Genspark, was founded by former Baidu executives Jing Kun and Zhu Kaihua in June 2024, initially focusing on AI search [8][32]. - The company raised $60 million (approximately 435 million RMB) in seed funding, achieving a post-money valuation of over 1.8 billion RMB [8]. - Genspark's initial positioning as an AI search product has shifted towards a broader focus on AI agents [35]. Group 3: Features and Capabilities - Genspark integrates various AI models, including DeepSeek V3 and R1, and utilizes a hybrid system with nine different scales of large language models (LLMs) [29]. - It offers over 80 tools and more than ten advanced datasets extracted from the internet [30]. - The agent has demonstrated its ability to create functional applications, such as a Flappy Bird-style game, in just two minutes [22]. Group 4: Market Context and Trends - The AI search market is becoming increasingly competitive, with significant resources and user attention concentrating on leading companies [44]. - The article notes that as AI agents gain popularity, entrepreneurs face critical decisions regarding their business strategies [46]. - Genspark's transition from AI search to a more versatile AI agent reflects broader industry trends and the need for adaptability in startups [42].

Core Viewpoint - The article emphasizes the transformation of AI tools, highlighting that the future will see AI as a tool used by humans, rather than merely a replacement for traditional search methods [4][5][14]. Group 1: Product Development and Features - The launch of "New Quark" represents a significant upgrade, introducing the "Super Box" concept, which aims to redefine the relationship between humans, tools (AI), and tasks [3][4]. - The "Super Box" is designed to be a "Super Agent" that directly delivers results by integrating various AI capabilities, such as AI search, writing, and health assistance, into a single interface [4][5]. - Quark's user base has grown significantly, with over 200 million monthly active users and a cumulative download exceeding 370 million by 2024, indicating strong market acceptance [8]. Group 2: Strategic Vision and Market Position - The strategic focus on Quark as a core component of Alibaba's AI To C strategy reflects a shift towards consumer-oriented AI products, aiming to provide comprehensive solutions for everyday tasks [6][7]. - The leadership of Wu Jia, who has extensive experience within Alibaba, is pivotal in driving Quark's growth and aligning it with the company's broader AI mission [9][10]. - The "Super Box" is positioned as a versatile tool that can cater to a wide range of user needs, from academic reports to travel plans, emphasizing its utility across various demographics [21][22]. Group 3: Future Directions and Innovations - Future developments will focus on enhancing the "Super Box" to support multi-modal inputs, allowing users to interact through various means such as voice and images [22][23]. - The integration of deep thinking capabilities and multi-modal abilities will enable the "Super Box" to handle more complex tasks, further distinguishing it from traditional search engines [35]. - The overarching goal is to establish Quark as a leading "Super Entrance" in the AI era, facilitating seamless interactions between users and information [40].

据介绍，知乎直答新版本重在强化AI与社区，尤其是与创作者的融合。除上传解析网页、文档等知识库之外，用户还可以 将每个知乎答主视为一个独立知识库，聚焦答主的单篇或全部创作进行提问、检索和解析，获得更具针对性的专业解答。 知乎直答新升级：溯源答主社区真实创作，强化AI与人的连接 知乎直答经过数次产品迭代，积累了知乎社区海量专业讨论、超5000万篇文献知识库的内容优势，并通过接入DeepSeek实 现推理能力升级。本次新版本则在内容和算法基础上，进一步挖掘人的专业价值，拓展了AI搜索的产品边界。 3月24日，知乎直答发布新版本，支持搜索结果直接溯源知乎答主真实创作，强化AI工具与人的连接，从而进一步降低AI幻 觉，提高内容的专业性和可信度。新版本的发布，意味着知乎直答与知乎社区的融合进入新阶段，有望为AI搜索产品和社区创 作生态带来更大发展空间。 同时，知乎直答的搜索结果支持直接溯源知乎答主的真实创作，打通AI工具与答主内容池，让AI生成结果实现源头可查、 专业可信。此举也将极大降低AI幻觉，为AI搜索的用户体验提供内容源头保障。 此外，知乎直答新版本将专业搜索和通用搜索两种模式合并，搜索结果一键调取所有知识库提供 ...

Investment Rating - The report maintains a "Buy" rating for the company, indicating a potential upside in the stock price [3][7]. Core Insights - The company reported a total revenue of 660.3 billion RMB for 2024, reflecting a year-over-year increase of 8%. The Non-IFRS net profit reached 222.7 billion RMB, up 41% year-over-year, while the profit attributable to equity holders was 194.1 billion RMB, marking a 68% increase [7]. - The company is expected to benefit from AI integration in its gaming and advertising sectors, with new AI-driven games and enhanced advertising platforms anticipated to drive growth [8]. - The company’s capital expenditure for 2024 was 76.7 billion RMB, a significant increase of 221% year-over-year, with plans to further increase spending in 2025 [9]. Company Overview - The company operates in the media industry, with a current H-share price of HK$540.00 and a market capitalization of approximately 378.76 billion RMB [2]. - Major shareholders include MIH Holdings B.V., which holds 24.43% of the shares [2]. - The company has a diversified product portfolio, with financial technology and enterprise services accounting for 31.3%, online gaming 30.1%, social networking 18.8%, and online advertising 18.5% [4]. Financial Performance - For Q4, the company achieved a revenue of 172.4 billion RMB, a year-over-year increase of 11%, and a quarter-over-quarter increase of 3%. Operating profit for the quarter was 51.5 billion RMB, up 24% year-over-year [7]. - The company’s net profit for 2025 is projected to be 221.6 billion RMB, representing a year-over-year growth of 14.18% [12]. Market Position - The company is positioned to leverage its stronghold in the AI search market, particularly through its WeChat platform, which has a daily search volume of nearly 600 million, compared to Baidu's 1.5 billion [9]. - The integration of AI is expected to enhance user engagement and advertising effectiveness, potentially leading to significant revenue growth in the advertising segment [9].

Investment Rating - The report maintains a "Buy" rating for the company, indicating a potential upside in the stock price [3][7]. Core Insights - The company reported a total revenue of RMB 660.3 billion for 2024, representing a year-over-year increase of 8%. The Non-IFRS net profit reached RMB 222.7 billion, up 41% year-over-year, while the profit attributable to equity holders was RMB 194.1 billion, reflecting a 68% increase [7]. - The company is expected to benefit from AI integration in its gaming and advertising sectors, with new AI-driven games and enhanced advertising platforms likely to drive growth [8][9]. - The company plans to increase capital expenditures in 2025, with a focus on AI-related research and development, and has set a target for share buybacks of at least HK$ 80 billion in 2025 [9]. Company Overview - The company operates in the media industry, with a market capitalization of approximately HK$ 378.76 billion and a current share price of HK$ 540.00 [2]. - Major shareholders include MIH Holdings B.V., which holds 24.43% of the shares [2]. - The company's stock has shown significant growth over the past year, with a 101.73% increase [2]. Financial Performance - For Q4, the company achieved a revenue of RMB 172.4 billion, a year-over-year increase of 11%, and a quarter-over-quarter increase of 3%. Operating profit for the quarter was RMB 51.5 billion, up 24% year-over-year [7]. - The company’s gaming revenue in Q4 grew by 23% to RMB 33.2 billion, driven by both existing popular games and new releases [9]. - The forecast for net profit attributable to equity holders for 2025 is RMB 221.6 billion, representing a year-over-year growth of 14.18% [12]. Product Mix - The company's revenue is diversified across several segments: Financial Technology and Enterprise Services (31.3%), Online Games (30.1%), Social Networks (18.8%), and Online Advertising (18.5%) [4].

Summary of Conference Call Notes Industry or Company Involved - The discussion revolves around the AI industry, specifically focusing on the developments and models from a company referred to as "Deep Sick" [1][2][3]. Core Points and Arguments 1. **Model Series Overview**: Deep Sick has released several models, notably V3 and R1, which are considered high-performance and cost-effective. The V3 model is highlighted for its engineering optimization and performance [1][2]. 2. **Comparison with Competitors**: The V3 model is compared to OpenAI's GPT-4, suggesting that it operates at a similar level of capability. The discussion emphasizes the importance of responsible AI development [2][3]. 3. **Scaling Laws**: The concept of "shifting on the curve" is introduced, indicating that as models evolve, they can achieve similar performance with fewer parameters, leading to cost reductions over time [3][4]. 4. **R1 Model Characteristics**: The R1 model is designed for long reasoning tasks, capable of handling complex queries. It has gained significant user engagement, reaching nearly 30 million monthly active users shortly after its release [5][6]. 5. **User Demographics**: Only 30% of R1's users are from China, indicating a strong international presence and appeal [6]. 6. **Innovative Training Approach**: The R1 model employs an object reward model (ORM) for training, which differs from traditional supervised fine-tuning methods, allowing for more flexible learning [7][8]. 7. **Consumer Applications**: The AI search capabilities of Deep Sick are highlighted as a rapidly growing application area, with the potential to provide reliable answers to user queries [10][11]. 8. **Market Impact**: The success of Deep Sick is seen as a catalyst for innovation in the AI sector, with implications for various industries, including healthcare and legal services [12][21]. 9. **Resource Requirements**: The discussion notes the significant computational resources required to support the models, with estimates suggesting the need for thousands of high-performance GPUs [19][20]. 10. **Future Outlook**: The potential for new applications and the overall positive sentiment towards the AI industry is emphasized, despite the presence of market bubbles [23]. Other Important but Possibly Overlooked Content 1. **Training Costs**: The narrative around the cost of developing AI models is nuanced, with claims that the reported costs may not fully capture the total investment required for development [16][17]. 2. **Externalities of Open Source**: The open-source nature of Deep Sick's models is seen as beneficial for fostering innovation and entrepreneurship within China [22][23]. 3. **Market Dynamics**: The call highlights the competitive landscape, noting that while some companies may struggle, others are likely to emerge successfully from the current market conditions [23]. This summary encapsulates the key insights and discussions from the conference call, providing a comprehensive overview of the current state and future potential of the AI industry as represented by Deep Sick.

AI搜索领域又一颗新星，成立不到两年估值翻至5.3亿美元。 其估值从之前的2.6亿美元增加到了现在的 5.3亿美元，投资方暂未披露。 据睿兽分析，成立于2023年底的MainFunc，2024年6月完成了 6000 万美金的种子轮融资，由蓝驰创投领 投。 详情请戳视频 以上就是本期全部内容，如果您想了解更多 最新 企业 动态， 欢迎登录 睿兽分析 查看，解锁 赛道图 谱 以及 行业报告 。 近日，前小度CEO景鲲和前小度CTO朱凯华联合创立的AI创新产品公司MainFunc，完成1亿美元的A轮融 资。旗下首款AI Agent搜索产品Genspark月活用户已经突破了200万。 ...

Core Insights - The article emphasizes the transition from traditional keyword-based search engines to AI-driven search models, highlighting the launch of Quark AI Search's "Deep Thinking" as a pivotal moment in redefining search engine capabilities [1][27] - Quark's "Deep Thinking" model utilizes reasoning to provide comprehensive answers, moving beyond simple information retrieval to a deeper understanding of user intent and context [2][13] Group 1: AI Search Evolution - Traditional search engines rely on keyword matching and webpage ranking, which often leads to fragmented information retrieval for users [2][23] - Quark's "Deep Thinking" model simulates human reasoning, allowing for multi-step inference and providing integrated, in-depth responses to complex queries [2][3] Group 2: Z Generation Social Product Design - The article outlines a systematic framework for designing social products tailored to Generation Z, focusing on their core characteristics such as individual expression, fragmented attention, strong privacy awareness, and emotional needs [4][6] - Key design principles include decentralized identity construction, low-pressure social environments, gamified experiences, and AI-enabled emotional connections [6][7] Group 3: Healthcare Information Retrieval - In the healthcare sector, Quark AI Search demonstrates its superiority by providing systematic solutions to queries, such as comprehensive advice on pollen allergy prevention, rather than just returning scattered links [9][10] - The model's ability to deliver structured health information showcases its professional advantages in the medical field [9][12] Group 4: Future of AI Search - Quark aims to evolve from a traditional search engine to an AI omnipotent assistant, capable of supporting various professional tasks such as report writing, research analysis, and strategic decision-making [23][24] - The anticipated integration of larger reasoning models indicates a significant expansion of Quark's capabilities, positioning it as a knowledge assistant and innovation catalyst in the future [25][27]

优秀的 AI 搜索不仅不能说谎、提供即时、全面的信息，更要足够 " 勤奋 " ，不断学习追踪、严格筛选那些真正有价值的资讯与答案，同时满足用户在某一 领域长期深耕、学习的诉求。 从某种意义上说，「 万物追踪」可以被视为AI时代的「今日头条」，但它又超越了传统的信息推送模式。通过AI技术的深度应用，它不仅能够提供个性化 的内容推荐，更能够根据用户的知识结构和学习需求，打造专属的信息获取体验。 Z Highlights 在小红书等社交平台上，如果你热衷于追踪各类前沿AI产品，那么 「 万物追踪」这个名字一定会频繁出现在你的视野中。这款由Minimax内部研发的新产 品，在2025年春节前夕正式进入公测阶段，曾在小红书上收获各种好评。 根据官方介绍，「 万物追踪」是一款基于大模型驱动的智能信息助手，其核心功能是通过持续监控和搜索互联网上的海量信息，智能识别与用户需求匹配 的内容，并经过专业化的提炼和总结，以精准推送的方式呈现给用户。 01 查询之上更进一步， AI 搜索前路如何？ 搜索引擎的出现一度被视为互联网技术发展的关键节点。然而，搜索引擎依旧偏重于让用户自己从海量的返回信息中挑拣适合的信息，此外，由于信息茧 ...