胡定坤 日前，美国众议院中国事务特别委员会主席约翰·穆勒纳尔致信美国福特汽车公司首席执行官吉姆·法 利，就福特与宁德时代等中国电池企业的合作情况进行质询。信中污蔑中国严重威胁美国供应链独立性 和经济安全，并表达了对中国"利用"美国关键产业和将基于中国技术所产电池部署在美国关键基础设施 中的风险等问题的担忧。 很明显，穆勒纳尔此次质询的目的就是对福特施压，阻挠其与中国电池企业的技术合作。2025年12月， 福特宣布将利用宁德时代的技术授权生产磷酸铁锂（LFP）电池，用于电力储能系统。今年1月15日， 美国《华尔街日报》披露福特还在与比亚迪洽谈混合动力汽车电池交易。当日，穆勒纳尔就发表声明 说，"这是一个严重的安全隐患"，福特应该与美国的盟友而不是对手合作。 讽刺的是，美国议员却无视本国企业利益，频繁发声、质询予以阻挠。殊不知福特将是合作推进的主要 受益者，也将是合作受阻的主要受害者。没有与福特的合作，宁德时代、比亚迪仍然是全球电池产业的 领军者。而与中企的合作拖延一天，福特就离拥有高性能LFP电池、离进军储能电池市场远了一天，美 国相关产业也必然因此受损。 美国议员做出如此不智之举的背后是"泛化国家安全"思维上 ...

(资料图) 运输机器人搭载着各类建材和小型设备来回穿梭，转运物资；焊接机器人在高空舞动"手臂"，自动 进行焊接作业；喷涂机器人在室内"挥毫泼墨"，自主完成墙面粉刷……这是记者在四川芯烨智创机器人 科技有限公司（以下简称芯烨机器人）举办的新闻发布会上看到的一幕。 易富民表示，智能建造机器人生态集群的发布，是企业拥抱行业变革，推进智能建造创新成果产品 化、产业化的起点。面向"十五五"，公司计划研制推出30款覆盖核心施工场景的建筑机器人产品，不断 增强企业核心功能和核心竞争力，加速构筑产业新生态。 （活动主办方供图） 标签： 最新资讯 本次发布的智能建造机器人生态集群，包含行业首创的项目级机器人集群智能调度管理系统和四款 自主研发的智能建造机器人。 芯烨机器人副总经理、首席技术官林宗奇介绍，机器人集群智能调度管理系统能够对各类机器人及 智能装备进行全生命周期管理与集群智能调度。该系统融合AI大模型技术，实现了智能巡检与视频监 控的深度融合，支持施工数据的实时3D可视化呈现与多屏同步查看，并具备异常自动报警功能，真正 实现施工全程透明可控。 同步推出的四款智能建造机器人，都是为解决建筑施工中的关键痛点而研发。高空焊接 ...

Group 1 - The core idea of the articles revolves around the emergence of Generative Engine Optimization (GEO) services, which are increasingly favored by businesses to manipulate AI-generated content for advertising purposes, raising concerns about compliance and consumer rights [1][2] - GEO services are characterized by their "deceptive" nature, packaging commercial promotions as objective AI-generated content, which may infringe on consumer rights as per advertising laws that require clear identification of advertisements [2] - The development of AI models relies on high-quality training data, and allowing commercialized, homogeneous information to contaminate the training sources could hinder technological evolution in the long run [3] Group 2 - The upcoming Central Economic Work Conference in 2025 emphasizes the need to "improve AI governance," highlighting the necessity for updated governance strategies that adapt to new technologies and scenarios [3] - There is a call for proactive regulatory measures that address the unique characteristics of new business models, including clarifying responsibilities for service providers and platforms to eliminate gray areas [3] - The challenges posed by AI advertising reflect broader governance issues in the face of rapid technological change, necessitating a collaborative approach that incorporates diverse expertise for effective governance [4]

Core Viewpoint - The construction of the Chongqi Railway and Highway Yangtze River Bridge and the Chongtai Yangtze River Tunnel is set to revolutionize high-speed rail travel in China, allowing trains to cross the Yangtze River at speeds of 350 km/h without slowing down, marking a significant advancement in infrastructure development [2][3][6]. Group 1: Bridge Construction - The Chongqi Railway and Highway Yangtze River Bridge features a main span of 400 meters, making it the longest of its kind under construction globally [3]. - The bridge employs a dual-layer design, accommodating a six-lane highway on the upper level and high-speed rail on the lower level, creating a three-dimensional transportation system [3][4]. - Engineers have overcome significant challenges related to bridge deformation caused by wind and train movement, ensuring the stability required for high-speed rail [3][4]. Group 2: Tunnel Construction - The Chongtai Yangtze River Tunnel is the largest high-speed rail underwater tunnel in China, with a total length exceeding 14 kilometers and a single-shield tunneling distance of 11.3255 kilometers, the longest in the world [5][6]. - The tunnel construction utilizes advanced technology, including a shield machine equipped with an intelligent system for autonomous operation, ensuring precision and safety under challenging conditions [6][7]. - Innovative materials and structural designs have been implemented to enhance the tunnel's stability and waterproofing, allowing it to withstand high water pressure and complex geological conditions [6][7].

1月28日，北京智源研究院主导的多模态大模型成果"Emu3"正式上线国际顶级学术期刊《自然》正刊（纸质版预计将于2月12日刊发），这是我国 科研机构主导的大模型成果首次登陆该期刊，标志着我国在人工智能原始创新领域取得重大突破。 此前，语言大模型依托"预测下一个词元（NTP）"的自回归路线实现重大突破，但多模态模型仍依赖对比学习、扩散模型等专用路线，自回归能 否成为多模态通用路线一直是行业未解之谜。智源团队提出的Emu3模型，将文本、图像、视频统一离散化到同一表示空间，基于单一Transformer 架构从零开始联合训练，仅凭"预测下一个词元"就实现了多模态生成与感知的统一。 实验显示，Emu3在文生图任务中性能比肩扩散模型，视觉语言理解能力媲美CLIP与大语言模型融合方案，还能以纯自回归方式生成高保真视 频，支持视频延展、图文交错生成及机器人操作建模等多元任务。《自然》编辑点评称，该成果对构建可扩展、统一的多模态智能系统具有重要 意义。 值得关注的是，该团队通过大规模消融实验验证了多模态学习的规模定律，证实直接偏好优化（DPO）可无缝适配自回归视觉生成。后续迭代的 Emu3.5更实现"预测下一个状态"的能力跃 ...

Core Insights - The AI industry in China is experiencing intensified competition as major players like Baidu, Alibaba, and DeepSeek rapidly innovate and release new technologies and products, establishing a three-way competitive landscape [1][9] Group 1: Baidu's Developments - Baidu launched the official version of its Wenxin 5.0 model on January 22, which utilizes a native multimodal architecture for unified modeling, supporting various data types such as text, images, audio, and video [1] - Wenxin 5.0 has achieved top rankings in both domestic and international arenas, being recognized as the leading model in text and visual understanding categories [1][2] - Baidu's new Paddle OCR-VL-1.5 model, released on January 29, features innovative "irregular frame positioning" technology for improved optical character recognition, directly competing with DeepSeek's offerings [4][8] Group 2: Alibaba's Innovations - Alibaba introduced the Qwen3-Max-Thinking reasoning model, which employs a novel testing time extension mechanism to enhance reasoning performance while being more cost-effective [3] - The model leverages its ecosystem by integrating with platforms like Taobao and Alipay, facilitating efficient collaboration between technology and application scenarios [3] Group 3: DeepSeek's Strategy - DeepSeek focuses on open-source advantages and bottom-layer capabilities, releasing the DeepSeek-OCR 2 model that utilizes the DeepEncoder V2 method for smarter image processing [3] - The company emphasizes a full-stack open-source approach, combining model weights, training frameworks, and deployment tools to maximize cost-effectiveness [3] Group 4: Industry Trends - The AI sector in China is transitioning into a new phase of large-scale implementation, with companies demonstrating clear paths of innovation and contributing to industry transformation and broader social benefits [9] - The competition among leading players is characterized by their unique business ecosystems and technological advancements, driving the industry from a "catch-up" phase to a "leading" position [9]

所有结构均表现出优异的机械稳定性。尽管颗粒间无化学键连接，但依靠强大的范德华力，它们仍能牢 固结合，实现自支撑。 最新研究有望彻底突破这一瓶颈。其核心创新在于，利用飞秒激光精准操控光流体效应，在液体中引导 纳米或微米级颗粒进行3D自组装。当激光聚焦于悬浮粒子溶液中某一点时，会瞬间加热形成"热点"， 引发局部热梯度，驱动周围流体产生定向流动——这种"光流"就像无形之手，将粒子推入预设的微模具 中。 团队表示，这项"光流组装"技术不仅克服了传统3D打印的材料限制，更打开了通往多功能微系统的大 门。未来，它有望应用于开发智能药物递送、体内微型手术机器人、高密度微传感器等前沿领域。 这种微模具如同一个微型"蛋糕模"，侧面设有小口。粒子随流而动，穿过缝隙，精准落入模腔，层层堆 积成所需形状。模具可设计为立方体、球体，甚至牛角面包等复杂曲面结构，自由度极高。组装完成 后，通过后处理去除聚合物模具，便留下一个完全由目标材料构成的独立微结构。 据新一期《自然》杂志刊发的研究，德国马克斯·普朗克智能系统研究所与新加坡国立大学科学家携 手，开发出一种利用光流体效应进行3D微纳制造的新技术。该技术不再局限于传统的聚合物，而是可 灵 ...

Core Insights - A research team led by researchers from the Guangzhou Institute of Geochemistry of the Chinese Academy of Sciences has utilized in-situ liquid-phase transmission electron microscopy to observe the dynamic process of gold nanoparticle formation on the surface of pyrite in a liquid environment for the first time [1][2] - The study proposes a new mechanism for gold precipitation induced by pyrite, revealing the presence of a special "dense liquid layer" at the pyrite-water interface, which acts as a "nano-factory" to catalyze the nucleation, growth, and enrichment of gold in extremely low concentration gold-bearing fluids [1][2] Research Findings - The research indicates that gold nanoparticles do not form spontaneously in the solution but are generated within the "dense liquid layer" closely adhering to the pyrite surface [2] - The dissolution of pyrite significantly reduces the oxygen fugacity within this layer, altering the local chemical environment and facilitating the rapid supersaturation of gold ions, leading to their precipitation as solid particles [2] - This mechanism can drive the migration and enrichment of gold at the interface even when the external solution has an extremely low gold concentration, challenging the traditional view that gold primarily originates from deep hydrothermal fluids [2] Implications - The findings provide new insights into the micro-dynamics of gold enrichment in both hydrothermal gold deposits and surface gold accumulation processes, opening new pathways for understanding mineralization processes driven by nanoparticles in nature [2] - The research also holds significant implications for interface regulation in green gold leaching processes [2]

Core Insights - The research team led by Zhang Bin at the Shanxi Coal Chemistry Research Institute has made significant progress in the hydrogenation of α-olefins, with results published in the international journal "ACS Catalysis" [1] Group 1: Research Findings - Synthesis gas generated from coal or biomass gasification can produce a range of α-olefins through iron-based Fischer-Tropsch synthesis, which can further react with synthesis gas via hydrogenation to yield long-chain fatty aldehydes, important raw materials for fragrances, pharmaceuticals, and surfactants [1] - The transition towards "refined and high-value" coal chemical processes is highlighted as a new direction for the industry [1] Group 2: Catalyst Development - The research team proposed a new approach to construct a "Co-MoC interface" by controlling the morphology and thickness of molybdenum carbide to optimize the interface structure [2] - Two efficient interface sites were developed: one is the Co-MoOx site formed by partially oxidized molybdenum carbide nanoparticles and cobalt particles, which significantly lowers the energy barrier for the hydrogenation reaction, achieving a catalytic activity 10.7 times higher than traditional cobalt-based catalysts [2] - The second site is the Co-Mo₂C site formed by molybdenum carbide nanosheets and cobalt atoms, which effectively addresses catalyst loss issues and greatly enhances reaction stability [2] Group 3: Implications for Industry - This research provides a new strategy for designing efficient and stable cobalt-based hydrogenation catalysts, offering a feasible technological route for the high-value conversion of coal resources [2] - The findings are expected to promote the green upgrade of the fine chemicals industry [2]

Core Insights - Shanxi Coking Coal Group is implementing intelligent transportation systems in its mines, significantly improving efficiency and safety in coal transportation [1][2] - The transition from traditional coal transportation methods to smart, automated systems is driven by the need to address inefficiencies and high energy consumption [1] Group 1: Intelligent Transportation Systems - The intelligent transportation system allows for remote control, fault monitoring, and modern management of coal transportation, reducing the need for manual labor [1] - The traditional coal transportation system faced challenges such as high energy consumption (60% of total energy) and safety risks due to reliance on manual labor [1] - Shanxi Coking Coal is focusing on optimizing the collaboration between people, machines, and the environment to enhance the efficiency of coal transportation [1] Group 2: Technological Advancements - The use of permanent magnet variable frequency direct drive technology has increased the efficiency of the driving system from 80% to over 94%, saving 51.11 million kWh annually and reducing maintenance frequency by 65% [2] - Intelligent inspection robots enhance the efficiency of the main transportation system by conducting comprehensive scans and generating real-time equipment health reports, reducing fault detection time by 69% [2] - The intelligent operation rate of the main transportation belts in Shanxi Coking Coal's mines has reached 74%, with an average reduction of 67% in labor requirements [2] Group 3: Strategic Goals - Shanxi Coking Coal aims to reshape the underground transportation ecosystem to achieve safe, efficient, and environmentally friendly production [2]