朋友们，我是帮主郑重。站在2026年的起点，我相信很多朋友和我一样，面对着看似矛盾的市场信号： 一边是机构预警AI泡沫可能破灭、全球增长面临不确定性；另一边，是量子计算、脑机接口这些颠覆 性技术加速向我们驶来。大家心里可能都有一个问号：2026年，机会到底在哪？今天的钱，又该投向何 方？ 作为一位和资本市场打了20年交道的观察者，我的答案不是一份"致富代码清单"，而是几条相对清晰 的、值得用中长线眼光去审视和布局的核心主线。在我看来，2026年的投资机会，正蕴藏在这"新与 旧"、"实与虚"、"内与外"的几重转换之中。 第二，坚持"深度研究"，警惕"故事投资"。 无论是AI还是其他赛道，随着产业成熟，市场会越来越"精 明"。仅靠宏大叙事已经不够，订单、收入、利润、现金流这些硬指标的重要性会空前提升。深入理解 一家公司的商业模式和竞争壁垒，比追逐热点标签更重要。 第三，保持"长期主义"的耐心与"中场休息"的纪律。 中长线投资不是买了就一动不动，它需要我们在 看准的长期方向上保持耐心，陪伴优秀公司成长。同时，也要有在中场（比如估值阶段性过高时）适度 休息、控制风险的纪律，这样才能走得更远。 总而言之，2026年的画 ...

↑点击上面图片 加雪球核心交流群 ↑ 风险提示：本文所提到的观点仅代表个人的意见，所涉及标的不作推荐，据此买卖，风险自负。 作者： 西峯 来源：雪球 一晃在论坛十五年了 ， 与二十多万朋友讨论交流 ， 深感荣幸 。 很多朋友常常问我同一个问题 ： " 怎么做好投资 ， 如何才能财务自由 ？ " 这个问题背后 ， 是所有人对未来的焦虑和期许 。 今天 ， 我想借这篇年度总结 ， 和大家聊聊过去20多年投资生涯中的一些想法 ， 希望能给大 家带来一些超越短期的启发 。 01 投资起点 ： 选对投资 " 大类标尺 " 我们谈论投资 ， 首先要有一个作为比较基准的锚 。 这个 " 锚 " ， 也就是你投资品所在 " 大类标尺 " （ Benchmark ） 。 它决定了获利的基 础概率 ， 也是投资者建立合理投资预期的前提 。 举个例子 。 我们看标普500指数 ， 在过去40年 ， 它的年化回报率大约在11.8%左右 ， 相当于40年85倍的水平 。 如果投资道琼斯指数 ， 它 的年化回报率只有9.3% ， 相当于40年35倍的收益 。 我们很容易意识到 ， 年化收益上的轻微差异 ， 经过40年的复利放大 ， 结果 ...

2015—2025年，东吴证券研究所实现从区域性研究部门到全市场最佳研究机构前十的跨越式发展。 以"机制+人才"双轮驱动，锚定苏州及长三角产业禀赋，深耕"产业研究+资本服务"差异化路径，构 建多元服务生态并强化内部协同。十年深耕下，研究业务已成为东吴证券发展的重要引擎。站在新 起点，东吴证券研究所将持续锚定深度研究与市场定价核心能力，全力冲刺国内一流券商研究所目 标。 12月17日，2025证券时报最佳分析师获奖名单在苏州正式揭晓，东吴证券研究所凭借扎实的研究实力，再度交 出亮眼答卷。在团队奖项方面，东吴证券研究所斩获含金量十足的最佳研究团队SSR（Superior Super Research）第七名，"最佳北交所公司研究团队"排名跃升至第三名，四大产业研究领域全部跻身前十。细分赛 道的表现同样可圈可点：不仅蝉联"汽车和汽车零部件"行业研究榜首，还在新能源和电力设备、传播与文化、 环保、非银金融等热门领域稳居前三，宏观、机械、医药等行业也成功跻身前五，展现出全面且强劲的研究竞 争力。 这一成绩的取得，恰逢东吴证券研究所转型卖方研究十周年之际，可谓生动印证了其十年深耕带来的实力跃 迁。 2015年，东吴证券研 ...

Core Insights - The article discusses the transformation of AI search from traditional search engines to native AI search, highlighting the competitive landscape among various startups and the need for a new search architecture focused on AI [1][3][5]. Group 1: Historical Context - In the 1990s, various startups explored different methods of internet search, with Yahoo using a directory approach and Google later revolutionizing the field with its PageRank algorithm [1][2]. - The emergence of Google in 1998 marked a significant shift, as its algorithm quickly became the preferred method for navigating the internet, effectively solving the search problem for users [2]. Group 2: Current Landscape - The current search environment is undergoing a major shift, with numerous startups competing to create AI-native search systems that can index the web for AI applications [3][6]. - Traditional web search is primarily optimized for human users, often resulting in cluttered results filled with ads and redundant information, which can hinder the effectiveness of AI models [3][5]. Group 3: Emerging Trends - The article posits that deep research will become a dominant and monetizable form of agent-based search, as clients are willing to pay for high-quality research outputs [5][17]. - Many companies are opting to outsource their search capabilities to specialized service providers due to the high costs and complexities associated with maintaining search infrastructure [7][15]. Group 4: Technological Innovations - New search architectures are being developed to support AI agents, focusing on real-time data access and dynamic information retrieval, which enhances the capabilities of AI models [11][12]. - The introduction of Retrieval-Augmented Generation (RAG) and Test-Time Computation (TTC) allows models to access real-time information and improve their reasoning capabilities, transforming static models into dynamic reasoning systems [11][12]. Group 5: Use Cases - Deep research has emerged as a prominent use case for AI search APIs, enabling agents to conduct extensive research tasks that would take humans significantly longer to complete [17][19]. - AI search is also being utilized for CRM lead enrichment, automating the process of gathering and updating relevant information from various sources [19]. - Real-time access to technical documentation and code examples is crucial for coding agents, ensuring they reference the most current and relevant information [20]. Group 6: Competitive Dynamics - The competitive landscape is shifting towards API platforms, where user-facing products can leverage various search functionalities through single integrations [15][22]. - Companies are increasingly evaluating search providers based on the quality of results, API performance, and cost, leading to a diverse range of offerings in the market [22][23].

Core Insights - The article highlights the development and capabilities of the domestic AI search product, Mita AI Search, which is one of the earliest AI search products launched by a startup team [1] - The collaboration with MiniMax Speech enhances the user experience by providing a natural and pleasant voice for knowledge explanations through the "Tazi Teacher" feature [2] - The M2 model from MiniMax is utilized for deep research, offering reasoning capabilities and a strong interlinked thinking chain [3] Group 1 - Mita AI Search is recognized for its excellent experience in knowledge indexing, organization, and output due to the team's background in computer science and law [1] - The "Deep Research" mode aims to demystify the search planning algorithm by presenting a dynamic "problem chain" during the analysis process, making complex research clearer [4] - MiniMax M2 is designed for top-tier coding and agentic capabilities, focusing on delivering excellent performance at optimal pricing [4]

Core Insights - The article discusses how to maximize the value of AI tools, emphasizing the importance of understanding user patterns and selecting the right AI model based on specific needs [1][3]. Group 1: AI Model Selection - Users have approximately nine choices for advanced AI systems, including Claude by Anthropic, Gemini by Google, ChatGPT by OpenAI, and Grok by xAI, with several free usage options available [3][4]. - For those considering paid accounts, starting with free versions of Anthropic, Google, or OpenAI is recommended before upgrading [4][6]. - The article highlights the differences in capabilities among AI models, such as web search efficiency, image creation, and handling complex tasks, which should guide user selection [4][7]. Group 2: Advanced AI Features - Advanced AI systems require monthly fees ranging from $20 to $200, depending on user needs, with the $20 tier suitable for most users [6][7]. - The article outlines the distinctions between chat models, agent models, and wizard models, recommending agent models for complex tasks due to their stability and performance [9][10]. - Users can choose specific models within systems like ChatGPT, Gemini, and Claude, with options for deeper thinking and extended capabilities [11][13][14]. Group 3: Enhancing AI Output - The article emphasizes the importance of "deep research" mode, which allows AI to conduct extensive web research before answering, significantly improving output quality [16][18]. - Connecting AI to personal data sources, such as emails and calendars, enhances its utility, particularly noted in Claude's capabilities [18]. - Multi-modal input options, including voice and image uploads, are available across various AI platforms, enhancing user interaction [19][20]. Group 4: Future Trends and User Engagement - The article predicts an increase in AI usage, with 10% of the global population currently using AI weekly, suggesting that user familiarity will evolve alongside model improvements [24]. - Users are encouraged to experiment with AI capabilities to develop an intuitive understanding of what these systems can achieve [24]. - The article warns against over-reliance on AI outputs, as even advanced models can produce errors, highlighting the need for critical engagement with AI responses [26].

Core Viewpoint - The article challenges the perception that brokerage asset management firms lack equity investment capabilities, highlighting that some firms, like Caitong Asset Management, have successfully established themselves in this area through active management and a focus on deep research and value investment [1][3]. Group 1: Company Overview - Caitong Asset Management has a total management scale exceeding 300 billion yuan, with nearly 110 billion yuan in public fund management, maintaining a leading position in the brokerage asset management industry [3]. - The firm has achieved a 156.69% absolute return rate for its equity funds over the past seven years, ranking in the top 20% among fund managers [3]. Group 2: Investment Philosophy and Team Structure - The investment philosophy of Caitong Asset Management is centered around "deep research, value investment, absolute returns, and long-term assessment," which has guided its equity investment strategy since its inception [4]. - The equity research team consists of approximately 40 members, with over 20 dedicated equity researchers, and has grown the scale of its equity public funds from 700 million yuan to over 17 billion yuan [4][5]. Group 3: Research and Investment Strategy - The firm has established a structured approach to integrate research and investment, with clear departmental divisions focusing on public and private equity investments, each led by experienced fund managers [8]. - Caitong Asset Management emphasizes a long-term investment strategy, focusing on fundamental research to uncover intrinsic value, regardless of market fluctuations [13][15]. Group 4: Team Development and Culture - The average experience of equity fund managers and investment managers at Caitong Asset Management exceeds 14 years, with many having backgrounds in absolute return investments [5]. - The firm fosters a culture of openness and shared values, encouraging diverse investment styles while ensuring that all team members receive adequate research support [12].

Core Viewpoint - The article emphasizes the resurgence of technological innovation in various sectors, highlighting the importance of deep industry research and understanding for successful investment in technology stocks [1][2]. Group 1: Importance of Deep Research - Continuous monitoring and deep immersion in the industry are essential for capturing opportunities behind changes in technology and market dynamics [2][3]. - The success of investment products from firms like Caitong Asset Management demonstrates the effectiveness of deep research and industry understanding [3][4]. Group 2: Investment Strategies and Performance - Caitong Asset Management's technology-focused funds have shown impressive performance, with specific funds ranking in the top 1% and 5% of their categories over various time frames [3][4]. - The investment strategies employed by fund managers like Bao Laiwen and Li Jing focus on understanding macro trends and industry dynamics, which have led to significant returns [7][10]. Group 3: Team Dynamics and Research Integration - The integration of research and investment processes is crucial, with a focus on building trust and collaboration between researchers and fund managers [15][18]. - The culture of sharing and collaboration within the team enhances the overall investment decision-making process, allowing for a more comprehensive understanding of market opportunities [22][23]. Group 4: Future Trends and Market Opportunities - The article discusses the potential of AI and other emerging technologies as key investment areas, with a focus on understanding real user needs and market demands [11][12]. - Caitong Asset Management's proactive approach to investing in AI-related sectors reflects a commitment to identifying and capitalizing on long-term industry trends [12][29]. Group 5: Performance Metrics - The performance metrics of Caitong Asset Management's funds indicate a strong track record, with specific annual growth rates and comparisons to benchmarks demonstrating effective management [30][31].

Core Viewpoint - The article discusses the new feature of Metaso AI's DeepResearch, highlighting its advanced capabilities in conducting in-depth research and analysis, particularly in the context of the competitive landscape of food delivery services in China. Group 1: Introduction of DeepResearch - Metaso AI has introduced a new feature called DeepResearch, which enhances its research capabilities beyond previous modes [5][6][7] - The author expresses a strong preference for using Metaso AI for research tasks, indicating a shift from other AI search products [3][4] Group 2: Functionality and User Experience - The new DeepResearch feature offers a game-like experience, making the research process engaging and intuitive [10] - The interface provides visual representations of the research process, including token usage, sources found, and time spent, enhancing user interaction [25][43] - The system allows for a comprehensive analysis of competitive dynamics, integrating both vertical and horizontal analyses of companies like JD, Meituan, and Taobao [18][19][20] Group 3: Research Output and Quality - The reports generated by DeepResearch are extensive, often exceeding 10,000 words, and are structured into clear chapters, providing detailed insights [52][60] - The analysis of the food delivery market reveals that the underlying cause of competition is "high frequency attacking low frequency," with Meituan being the primary aggressor [54][55][59] - The quality of the reports is noted to be comparable to that of OpenAI's DeepResearch, with precise and relevant findings [48][60] Group 4: Additional Features and User Control - Users can generate interactive visual reports, catering to preferences for visual data representation [66] - The platform allows users to manage source preferences, enhancing the customization of research outputs [67] - Metaso AI offers a generous daily search quota, making it accessible for frequent use compared to other paid services [69]

Core Insights - The article discusses the emergence and evolution of "Deep Research" systems in the AI Agent exploration wave, highlighting the rapid development and competition among major players like Google, OpenAI, and Anthropic since late 2024 [1][2] - A comprehensive survey from Zhejiang University provides a framework for understanding and evaluating the current landscape of deep research systems, emphasizing the shift from model capability to system architecture and application adaptability as the main competitive focus [1][2] Group 1: Current Landscape and System Comparisons - The ecosystem of deep research systems is characterized by significant diversity, with different systems focusing on various technical implementations, design philosophies, and target applications [3] - Key differences among systems are evident in their foundational models and reasoning efficiency, with commercial giants leveraging proprietary models for superior performance in handling complex reasoning tasks [4] - Systems also differ in tool integration and environmental adaptability, showcasing a spectrum from comprehensive platforms to specialized tools [5] Group 2: Application Scenarios and Performance Metrics - In academic research, systems like OpenAI/DeepResearch excel due to their rigorous citation and methodology analysis capabilities, while in enterprise decision-making, systems like Gemini/DeepResearch thrive on data integration and actionable insights [8] - Performance metrics reveal that leading commercial systems maintain an edge in complex cognitive ability benchmarks, although specialized evaluations highlight the strengths of various systems in specific tasks [9][10] Group 3: Implementation Challenges and Technical Solutions - The implementation of deep research systems involves strategic trade-offs across architecture design, operational efficiency, and functional integration [12] - Core challenges include managing hallucination control, privacy protection, and ensuring interpretability, with solutions focusing on source grounding, data isolation, and transparent reasoning processes [15] Group 4: Evaluation Frameworks - The evaluation of deep research systems is evolving from single metrics to a multi-dimensional framework that assesses functionality, performance, and contextual applicability [16] - Functional evaluations focus on task completion capabilities and information retrieval quality, while non-functional assessments consider performance efficiency and user experience [17][18] Group 5: Future Directions in Reasoning Architecture - Future advancements in deep research systems are expected to address limitations in context window size, enabling more comprehensive analysis of large-scale research materials [22][23] - The integration of causal reasoning capabilities and advanced uncertainty modeling will enhance the systems' applicability in complex fields like medicine and social sciences [27][30] - The development of hybrid architectures that combine neural networks with symbolic reasoning is anticipated to improve reliability and interpretability [25][26]