2025.12. 17 德勤中国高科技高成长项目全国主管合伙人赵锦东表示，在企业推动科技与创新时，人才、资本与 AI研发投入成为三大关键驱动力。调研显示，23%的中国50强企业和66%的明日之星企业，AI研发 投入占营收的50%以上。但无论是中国50强企业还是明日之星企业，都面临着高技术人才的短缺、 AI技术在业务场景中的应用不足以及研发成本攀升等挑战。然而，核心技术自主研发、产品的快速 迭代和多元化融资正成为企业韧性发展的突破口。此外，越来越多的科技企业将AI智能体从工具化 向数字员工化演进，以全面提升效率、创新力与竞争优势。 全球科技行业正经历一场由AI驱动的深层次变革，算力主权竞争、开源模型生态崛起、软硬件融合 下的具身智能落地、AI智能体进化，以及AI全域化赋能，共同构成当前AI科技的核心发展趋势。"在 这趋势的带动下，中国科技行业的高潜赛道已向多领域拓展布局，企业凭借技术创新与生态化场景落 地，在创新提质中抢占发展高地。"赵锦东说，2025~2030年，中国将进入"AI+制造/新能源/生命"多 技术矩阵爆发期，成为全球"算力替代人力"的受益方与备份提供方。目前，中国科技正在从AI全面渗 透、算力与连 ...

在企业推动科技与创新时，人才、资本与AI研发投入成为三大关键驱动力。 德勤中国高科技高成长项目全国主管合伙人赵锦东表示，在企业推动科技与创新时，人才、资本与AI 研发投入成为三大关键驱动力。调研显示，23%的中国50强企业和66%的明日之星企业，AI研发投入占 营收的50%以上。但无论是中国50强企业还是明日之星企业，都面临着高技术人才的短缺、AI技术在业 务场景中的应用不足以及研发成本攀升等挑战。然而，核心技术自主研发、产品的快速迭代和多元化融 资正成为企业韧性发展的突破口。此外，越来越多的科技企业将AI智能体从工具化向数字员工化演 进，以全面提升效率、创新力与竞争优势。 全球科技行业正经历一场由AI驱动的深层次变革，算力主权竞争、开源模型生态崛起、软硬件融合下 的具身智能落地、AI智能体进化，以及AI全域化赋能，共同构成当前AI科技的核心发展趋势。"在这趋 势的带动下，中国科技行业的高潜赛道已向多领域拓展布局，企业凭借技术创新与生态化场景落地，在 创新提质中抢占发展高地。"赵锦东说，2025~2030年，中国将进入"AI+制造/新能源/生命"多技术矩阵 爆发期，成为全球"算力替代人力"的受益方与备份提供方 ...

Core Viewpoint - The article explores the application and evaluation of AI products in real-world scenarios, specifically focusing on how AI can enhance content creation and analysis for WeChat public accounts [1]. Evaluation Design - The evaluation faced typical performance bottlenecks of large language models (LLMs), particularly in large-scale data processing and rendering [3][4]. - The goal is to determine if AI can effectively analyze and provide insights into the content strategies of leading tech accounts [5]. Methodology - A "Hybrid Pipeline" approach was adopted, consisting of two phases: - Phase One: Python handles all quantifiable analysis tasks, producing a structured summary in JSON format [7]. - Phase Two: LLMs analyze the JSON data to generate reports, combining analytical rigor with AI insights [8]. Analysis Goals - The analysis aims to compute key metrics from WeChat articles, including topic distribution, posting rhythm, interaction metrics, and title style recognition [12][13]. Evaluation Process - The evaluation highlighted differences in performance among three models (Claude, Minimax, and Step 3) in code generation and file parsing [25]. - Claude and Minimax were chosen for their superior long-context architecture and multi-format file parsing capabilities [25]. Evaluation Results - The analysis of the three leading tech accounts (Quantum Bit, Founder Park, and Digital Life Kazk) revealed insights into their content strategies, including topic selection, posting frequency, and interaction structures [49]. - Key themes identified include the phenomenon of DeepSeek, AI agent applications, open-source model ecosystems, and the dynamics of industry giants like OpenAI [52][54]. Insights and Recommendations - Claude and Minimax suggested a balance between "traffic" and "depth" to enhance brand influence, noting the "efficiency paradox" where higher readership often correlates with lower engagement metrics [27]. - Recommendations include focusing on high-performing topics, optimizing posting times, and maintaining a rational narrative style while incorporating timely elements to enhance engagement [29][40]. Conclusion - The analysis concluded that successful accounts utilize data to identify topic potential, control content structure, establish professional trust through verification, and manage audience expectations through rhythm [72][74].

Core Insights - Alibaba's new model Qwen 3 is emerging as a leading force in the Chinese open-source AI ecosystem, replacing previous models like Llama and Mistral [1] - The interview with industry representatives highlights the importance of model fine-tuning, the choice between open-source and closed-source models, and the challenges faced in large model entrepreneurship [1] Model Selection - The majority of the company's needs (over 90%) require fine-tuned models for local deployment, with specific tasks utilizing APIs from models like GPT and Qwen [3] - Commonly used model sizes include 7B, 32B, and 72B, with smaller models (0.5B, 1.5B) for privacy-sensitive applications [3] - Qwen is preferred due to its mature and stable ecosystem, including well-adapted inference frameworks and fine-tuning tools [4] Technical Considerations - Qwen's strong support for Chinese language and its relevant pre-training data make it suitable for the company's focus on emotional companionship and psychological applications [6][7] - The complete series of model sizes offered by Qwen allows for lower fine-tuning costs and easier testing across different model sizes [7] Challenges in Model Usage - In embodied intelligence, challenges include high inference costs and ecosystem compatibility, especially when considering local deployment for privacy [9][10] - Online business faces challenges in model capability and inference costs, particularly during peak usage times [12] Model Capability and Business Needs - Current models do not fully meet the company's needs for nuanced emotional understanding, necessitating post-training to align models with specific business requirements [13] - The goal is to maintain general capabilities while significantly enhancing core domain abilities, with an acceptable trade-off in general performance [13] Open-source Model Development - The expectation is for open-source models to catch up with top closed-source models, with a desire for more technical details to be shared by developers [14] - Qwen and Llama focus on community and general usability, while DeepSeek is more aggressive in exploring cutting-edge technologies [15][16] Entrepreneurial Insights - A significant oversight in AI entrepreneurship is the mismatch between models and product needs, emphasizing the importance of understanding user requirements [17] - The correct approach is to integrate AI as a backend capability rather than a front-end interface, ensuring deeper personalization in user interactions [19] Global Impact of Open-source Models - The rise of Chinese open-source models like Qwen and DeepSeek is accelerating a global technological evolution, providing a path for Chinese companies to innovate and collaborate internationally [20]