Investment Rating - The industry investment rating is "Outperform the Market" with an expected increase of over 10% relative to the CSI 300 index in the next six months [41]. Core Insights - The report emphasizes that the introduction of sparse attention mechanisms, particularly through DeepSeek, significantly enhances the development potential of the AI industry [8][37]. - DeepSeek's advancements in attention mechanisms, including Native Sparse Attention (NSA) and DeepSeek Sparse Attention (DSA), are pivotal in improving model performance and efficiency [18][23][37]. Summary by Sections 1. Relationship Between Attention Mechanism and Large Model Development - The attention mechanism, introduced to improve information processing efficiency, has become a core component of large models, addressing the limitations of traditional recurrent neural networks [11]. - Sparse attention reduces computational complexity from O(L²) to sub-quadratic levels, thus overcoming memory and computational bottlenecks [11]. 2. DeepSeek's Technological Improvements in Attention Mechanism - DeepSeek has made significant contributions in three main areas: Multi-head Latent Attention (MLA), Native Sparse Attention (NSA), and DeepSeek Sparse Attention (DSA) [12][18][23]. - MLA reduces memory usage by approximately 90% while maintaining model performance, significantly lowering training costs [16]. - NSA enhances long text processing speed by 11 times and achieves performance comparable to traditional models [18]. - DSA improves training and inference efficiency, leading to substantial cost reductions for model usage [23]. 3. DSA and NSA Unlock Greater Development Potential for the AI Industry - The integration of DSA and NSA allows for expanded model context and improved computational efficiency, which are crucial for meeting the demands of multi-modal applications [33][37]. - The trend towards longer input and output lengths necessitates innovative approaches to model training and performance enhancement [33].

Core Insights - The article highlights the upcoming Baidu World 2025 event scheduled for November 13, 2023, in Beijing, focusing on three main areas: international expansion, AI applications, and advancements in large models [3][5]. Group 1: International Expansion - The trend of Chinese tech companies going global is gaining momentum, driven by AI technologies such as large models, intelligent agents, and autonomous driving [4]. - Baidu's autonomous driving service, Apollo Go, has achieved significant milestones in international markets, partnering with Uber and Lyft, and providing over 14 million rides globally, making it the leader in the sector [4][5]. Group 2: AI Applications - AI applications will be a central theme at Baidu World 2025, with the company continuing to enhance its internal products like search and cloud services, and launching the GenFlow universal intelligent agent [5][6]. - The new generation of digital human technology has shown commercial success, with a recent digital human broadcast generating 55 million yuan in GMV, and a new video generation model reducing costs by 70% [5][6]. Group 3: Advancements in Large Models - The event is expected to showcase upgrades to Baidu's Wenxin large model family, which has received recognition for its performance, with the latest Wenxin X1.1 model outperforming competitors like Deepseek-R1 [6]. - The Wenxin 4.5 series has been noted for its potential to solidify China's position in AI, recently topping the HuggingFace leaderboard, indicating strong international competitiveness [6][7].

Core Viewpoint - The strategic cooperation agreement between SenseTime and Cambricon aims to enhance the optimization of software and hardware, fostering an open and win-win industrial ecosystem in the artificial intelligence sector [2][4]. Group 1: Strategic Cooperation - SenseTime and Cambricon will leverage their respective technological and industrial resource advantages to develop domestic AI infrastructure, explore vertical business opportunities, and promote technology exports [2][3]. - The collaboration aligns with the national "AI+" strategy, combining SenseTime's strengths in large model development and AI infrastructure with Cambricon's expertise in intelligent computing chips [2][3]. Group 2: Market Response - Following the announcement of the partnership, SenseTime's stock surged over 5%, while Cambricon's shares increased by nearly 4% [1][2]. - The rapid response in stock prices indicates strong market sentiment towards the collaboration and its potential impact on the AI industry [1][2]. Group 3: Product Development - The two companies will focus on adapting the latest hardware and software products to create service solutions for the computing power market [3]. - They will also develop integrated solutions targeting vertical industry scenarios, enhancing their combined software and hardware capabilities [3]. Group 4: Industry Trends - The current phase of the tech stock market is characterized by significant volatility, with analysts suggesting that the market is in the early stages of a potential explosive growth phase [4][5]. - The ongoing trade tensions have reinforced the logic of domestic substitution, particularly in software, as companies seek to mitigate risks associated with foreign dependencies [4][5].

Investment Rating - The report maintains a positive outlook on the computer industry, with a focus on domestic AI applications and edge models [3][9]. Core Insights - DeepSeek's V3.2-Exp enhances long-text efficiency through DSA fine-grained sparse attention, achieving significant improvements in training and inference efficiency while reducing API costs by over 50% [3][9]. - OpenAI's launch of Sora 2 and Apps SDK outlines a business strategy focused on "entry + ecosystem," with Sora 2 allowing users to create AI-generated short videos, achieving over 1 million downloads in its first week despite being invite-only [3][9]. - Figure 03, featuring the Helix architecture, integrates cognition and action in humanoid robots, designed for mass production with a target of 120,000 units over four years, which will drive demand for related technologies [3][9]. Summary by Sections - **DeepSeek V3.2-Exp**: The release focuses on improving long-text processing efficiency with a new attention mechanism, significantly lowering operational costs for developers [3][9]. - **OpenAI Developments**: The introduction of Sora 2 and Apps SDK enhances user interaction with large language models, potentially increasing token consumption and creating new AI application opportunities [3][9]. - **Figure 03 and Helix Architecture**: The humanoid robot's design aims for high precision in tasks, with mass production capabilities that will stimulate demand across various sectors [3][9]. Key Targets - Key investment targets include Wuxi Unicomp Technology Co., Ltd., Kingdee International Software Group, Iflytek, Newland Digital Technology, Autel Intelligent Technology, Hand Enterprise, ArcSoft Corporation, and Hygon Information Technology Co., Ltd. [3][9].

Core Insights - Shanghai's Economic and Information Technology Commission has issued the "Action Plan for High-Quality Development of the Intelligent Terminal Industry (2026-2027)", emphasizing the enhancement of robotic terminal capabilities and the development of humanoid robots with emotional and cognitive skills [1] - The humanoid robot industry is entering a phase of rapid commercialization, with significant advancements in technology and increased participation from both domestic and international players [1][2] - The recent launch of Figure03 by FigureAI marks a significant step towards general intelligence in robotics, featuring upgraded perception systems and dexterous hands, indicating a shift towards mass production capabilities [2] Group 1 - The action plan aims to support the research and mass production of humanoid robots, focusing on core components like edge chips, dexterous hands, and batteries [1] - The emergence of AI companies like DeepSeek is driving the development of general-purpose humanoid models, leading to a diverse and competitive landscape in the humanoid robot industry [1] - The commercial viability of humanoid robots is becoming increasingly evident, with industrial applications gaining traction both domestically and internationally [1] Group 2 - FigureAI's Figure03 can autonomously handle household tasks such as laundry, cleaning, and dishwashing, showcasing advancements in sensory systems and dexterous manipulation [2] - The production capacity for Figure03 is projected to reach 10,000 units annually within four years, indicating a robust manufacturing strategy that moves away from CNC processing to more efficient methods [2] - The humanoid robot industry is expected to officially enter commercialization by 2026, with a focus on identifying high-quality companies within the supply chain for long-term investment opportunities [2]

Core Insights - The report highlights the significant growth in the automotive industry, with production and sales reaching 24.33 million and 24.36 million units respectively from January to September, marking a year-on-year increase of 13.3% and 12.9% [5][8] - The report emphasizes the positive performance of the financial and liquor sectors in the A-share market, indicating a potential for investment opportunities in these areas [5][9] - The gaming sector is projected to perform well due to favorable policies and AI-driven advancements, with a notable increase in revenue and profit for gaming companies [27][29] Domestic Market Performance - The Shanghai Composite Index closed at 3,865.23, down 0.62%, while the Shenzhen Component Index closed at 12,895.11, down 2.54% [3] - The A-share market is experiencing a period of consolidation, with significant trading volumes indicating investor interest [5][9] International Market Performance - The Dow Jones closed at 30,772.79, down 0.67%, and the S&P 500 closed at 3,801.78, down 0.45%, reflecting a general downturn in major international indices [4] Industry Analysis - The basic chemical industry showed a slight increase in revenue and profit in the first half of 2025, with total revenue reaching 1.300467 trillion yuan, a year-on-year growth of 4.7% [20][21] - The gaming industry is experiencing robust growth, with a nearly 24% increase in revenue and a 75% increase in net profit year-on-year [29][27] - The photovoltaic industry is facing challenges with a significant decline in new installations, down 55.29% year-on-year in August [23][24] Investment Recommendations - The report suggests focusing on investment opportunities in the soft drink, health products, and snack sectors, highlighting specific companies for potential investment [19][27] - In the gaming sector, the report recommends monitoring companies with strong product cycles and performance metrics, as well as those leveraging AI technologies [29][27]

Group 1 - The Shanghai Municipal Economic and Information Commission has issued the "Action Plan for High-Quality Development of the Intelligent Terminal Industry (2026-2027)", emphasizing the enhancement of robotic terminal capabilities and the development of humanoid robots with emotional intelligence and skills [1] - The report highlights a surge in domestic and international industry catalysts, with an increase in participants in the humanoid robot sector, and companies like Tesla and Figure AI accelerating their commercialization efforts [1] - The emergence of AI companies such as DeepSeek is driving the development of general-purpose robotic models, indicating a vibrant and competitive humanoid robot industry, with a clear trend towards industrial applications [1] Group 2 - Figure AI has officially launched Figure03, which can autonomously handle household tasks like laundry and cleaning, featuring upgrades in its perception system and dexterous hands [2] - The company has shifted its manufacturing approach from CNC processing to mold/injection/pressing techniques, with a production capacity of 12,000 units per year for the first generation and a target of 100,000 units over the next four years [2] - The humanoid robot industry is experiencing significant advancements, with a focus on short-term event-driven industry fluctuations and long-term attention on quality companies within the supply chain [2]

Investment Rating - The report maintains an "Outperform" rating for the computer industry, indicating an expected increase of over 10% relative to the CSI 300 index in the next six months [41]. Core Insights - DeepSeek has gained significant attention since the release of its R1 model earlier this year, and it has since focused on incremental updates rather than launching a more advanced R2 model. The development is categorized into three main stages: performance enhancement, hybrid reasoning architecture implementation, and cost reduction with accelerated domestic adaptation [7][10]. - The introduction of the V3.2-Exp model has led to a substantial reduction in API calling prices, with input cache hit prices dropping to 20% of R1's cost and output prices to 19%, enhancing the model's cost-effectiveness and market competitiveness [33][34]. Summary by Sections Stage One: Performance Enhancement - In March, DeepSeek launched V3-0324 and in May, R1-0528, which improved model capabilities through post-training, bridging the gap with leading models [11][12]. Stage Two: Hybrid Reasoning Architecture and Agent Capability Enhancement - From August onwards, DeepSeek aligned with global trends by releasing V3.1 and V3.1-Terminus, significantly enhancing agent capabilities and reasoning efficiency through extensive training on the DeepSeek-V3.1-Base model [12][18]. Stage Three: Efficiency Improvement and Domestic Adaptation Acceleration - The V3.2-Exp model, released in September, introduced a new attention mechanism (DSA) that improved training and reasoning efficiency while significantly lowering costs. This model also marked a milestone in the domestic AI industry, achieving zero-day adaptation with domestic chips from Huawei and Cambrian [31][34].

Core Insights - The value of scarce and certain assets is increasingly highlighted amid rising global uncertainties, including risks of a U.S. government shutdown and potential political changes in Japan [1] - Emerging markets are becoming a key direction for risk diversification, with Hong Kong positioned as a core platform for international capital allocation due to its market openness, high liquidity, and low correlation with U.S. dollar assets [1] - The technology narrative and AI wave are providing strong momentum for the growth of Hong Kong stocks, with significant increases in AI capital expenditure expectations following new model releases from institutions like OpenAI and DeepSeek [1] - The CSI Hong Kong Stock Connect Technology Index has risen over 51% year-to-date, with approximately 20% gains in August and September, reflecting the impact of the AI trend on Hong Kong stock trading [1] - The Dongwu Overseas Strategy Team notes that while short-term fluctuations in Hong Kong stocks are expected, they remain in an upward channel with limited downside risk, and potential Fed rate cuts may drive global capital towards equity assets, particularly benefiting the Hong Kong technology growth sector [1] Hong Kong Stock ETFs - The Hong Kong Stock Connect Technology ETF (159101) covers the entire technology industry chain [1] - The Hang Seng Internet ETF (513330) focuses on leading internet companies [1]

公众号记得加星标⭐️，第一时间看推送不会错过。 来源 ：内容 编译自 Wired 。 1999年，《过山车大亨》（Rollercoaster Tycoon）或许算不上最火爆的电脑游戏。但如果你深入探 究像素之下——摇摇晃晃的游乐设施，饥肠辘辘、口渴难耐、呕吐不止的人群（以及在他们身后拖地 的清洁工）——深入代码层面，你会看到这款游戏对工艺的执着近乎疯狂。游戏的唯一开发者克里斯 ·索耶（Chris Sawyer）用汇编语言编写了整款游戏。 某些编程语言，例如 Python、Go 或 C++，之所以被称为"高级"语言，是因为它们的工作方式有点 像人类语言，用命令和习语编写，这些命令和习语在诗歌朗诵会上或许会用得上。一般来说，像编译 器这样的软件会将这些语言转换成机器真正能读懂的东西：由 1 和 0（或者可能是十六进制）组成的 数据块，用来告诉实际的晶体管如何工作。汇编语言是"低级"语言中最低级的一种，它与机器的母语 几乎一一对应。它直接在机器上编码。用汇编语言开发一款复杂的电脑游戏就像用脱落的猫毛编织挂 毯。 为什么有人会这么做？效率是原因之一。在 20 世纪 90 年代，高级编程的工具并不齐全。编译器非 常慢。 ...