单词学习：![[cebd83c1-5ac2-4390-b427-c19c9a09b58d.mp3]]

新闻播报：![[065bb751-6e39-4efc-b1a0-f8c506162aac.mp3]]

工信部规划“人工智能+制造”蓝图，主攻芯片与大模型

简报：

• 2026年2月3日，工信部在中国人工智能产业发展联盟会议上表示，将全力推动“人工智能+制造”向深层次发展，将制造业打造为AI赋能的主战场。
• 为实现此目标，工信部明确了四项关键举措：重点突破算力芯片、工业大模型等核心技术；通过标准引领和开源文化优化应用环境；利用多边机制深化国际合作；以及加强行业自律与安全治理。
• 工信部科技司副司长甘小斌希望联盟能汇聚行业合力，构建具国际竞争力的产业生态。

相关链接：

• 工信部发力：突破算力芯片与工业大模型，绘就“人工智能+制造”新蓝图
• 挑战英伟达!英特尔 CEO 陈立武宣布进军 GPU 生产，发力 AI 算力市场

MIIT Plans "AI + Manufacturing" Blueprint, Focusing on Chips and Large Models

Brief:

• On February 3, 2026, the Ministry of Industry and Information Technology (MIIT) announced at the China AI Industry Alliance meeting its full commitment to advancing "AI + Manufacturing" to deeper levels, establishing manufacturing as the primary battleground for AI empowerment. 🚀
• To achieve this goal, MIIT has outlined four key measures: focusing on breakthroughs in core technologies such as computing power chips and industrial large models; optimizing the application environment through standard leadership and open-source culture; deepening international cooperation using multilateral mechanisms; and strengthening industry self-regulation and security governance. 💡
• Gan Xiaobin, Deputy Director of MIIT's Department of Science and Technology, expressed hope that the alliance would pool industry efforts to build an internationally competitive industrial ecosystem. 🌍

Related Links:

• MIIT's Efforts: Breaking Through Computing Power Chips and Industrial Large Models, Charting a New "AI + Manufacturing" Blueprint
• Challenging NVIDIA! Intel CEO Pat Gelsinger Announces Entry into GPU Production, Boosting AI Computing Power Market

鲁信创投与工银投资联手，10亿AI基金在山东落地

简报：

• 近日，“山东省鲁信工融新动能人工智能创业投资基金合伙企业（有限合伙）”正式成立，出资额达10亿元人民币。
• 该基金由鲁信创业投资集团股份有限公司、工银金融资产投资有限公司、山东省新动能绿色先行投资中心（有限合伙）等共同出资。
• 基金执行事务合伙人为工银资本管理有限公司和山东省高新技术创业投资有限公司，经营范围主要为私募股权投资及创业投资基金管理服务。

相关链接：

• 鲁信创投、工银投资等出资10亿元成立人工智能创业投资基金
• 武汉将启动建设超大规模 AI 之城，到 2028 年相关产业规模超 2000 亿元

Luxin Venture Capital and ICBC Investment Join Forces, RMB 1 Billion AI Fund Launched in Shandong ✨

Briefing:

• Recently, the "Shandong Luxin Gongrong New Kinetic Energy Artificial Intelligence Venture Capital Investment Partnership (Limited Partnership)" was officially established, with a capital contribution reaching RMB 1 billion. 💰
• The fund is jointly financed by Luxin Venture Capital Group Co., Ltd., ICBC Financial Asset Investment Co., Ltd., Shandong New Kinetic Energy Green Pioneer Investment Center (Limited Partnership), and other entities.
• The fund's executive partners are ICBC Capital Management Co., Ltd. and Shandong Hi-tech Venture Capital Co., Ltd. Its business scope primarily covers private equity investment and venture capital fund management services. 🚀

Related Links:

• Luxin Venture Capital, ICBC Investment, etc. Contribute RMB 1 Billion to Establish AI Venture Capital Fund
• Wuhan to Launch Construction of Ultra-Large-Scale AI City, with Related Industry Scale Exceeding RMB 200 Billion by 2028

中文大模型最新排名：海外领跑，国产局部登顶

简报：

• SuperCLUE发布的“2025年度中文大模型基准测评报告”显示，海外闭源模型Anthropic的Claude-Opus-4.5-Reasoning、谷歌的Gemini-3-Pro-Preview与OpenAI的GPT-5.2（high）包揽综合排名前三。
• 国产大模型在细分领域实现超越，其中Kimi-K2.5-Thinking在代码生成任务中夺得全球第一，Qwen3-Max-Thinking在数学推理上与谷歌并列世界冠军。
• 在开源领域，国产模型占据主导地位，实力大幅领先海外同类模型；智谱AI的开源模型GLM-4.7-Flash发布两周下载量突破100万，并在同尺寸模型中取得SOTA（State-of-the-Art）分数。

相关链接：

• 14天突破百万下载!智谱 GLM-4.7-Flash 领跑开源大模型 SOTA
• 2025 全球中文大模型榜单出炉：海外巨头包揽前三，国产大模型在细分领域超车

Latest Ranking of Chinese LLMs: Overseas Leaders, Domestic Models Achieve Local Supremacy 🚀

Brief:

• The '2025 Annual Chinese Large Language Model Benchmark Report' released by SuperCLUE indicates that overseas closed-source models — Anthropic's Claude-Opus-4.5-Reasoning, Google's Gemini-3-Pro-Preview, and OpenAI's GPT-5.2 (high) — secured the top three spots in the overall ranking.
• Domestic LLMs achieved breakthroughs in specific sub-domains, with Kimi-K2.5-Thinking claiming the global top spot in code generation tasks, and Qwen3-Max-Thinking tying with Google for the world championship in mathematical reasoning. 🏆
• In the open-source sector, domestic models dominate, significantly outperforming their overseas counterparts. Zhipu AI's open-source model GLM-4.7-Flash surpassed 1 million downloads within two weeks of its release and achieved SOTA (State-of-the-Art) scores among models of the same size. ✨

Related Links:

• Over 1 Million Downloads in 14 Days! Zhipu GLM-4.7-Flash Leads Open-Source LLMs to SOTA
• 2025 Global Chinese LLM Ranking Released: Overseas Giants Claim Top Three, Domestic LLMs Overtake in Specific Segments

昆仑万维发布桌面AI助手，主打本地文件处理能力

简报：

• 昆仑万维于2月4日面向全球发布了桌面端AI应用“天工 Skywork 桌面版”（SkyworkDesktop）。
• 该应用的核心特点是直接在用户电脑本地执行任务，能够读取、汇总和整理海量本地文件（如文档、图片、视频等），并生成新内容，全程无需上传文件至云端。
• 产品定位为“OS助手”和“Windows版Claude Cowork”，旨在融入桌面办公场景，处理跨文件、跨格式的复杂任务。
• 应用支持在Claude、Gemini等多种大模型之间进行选择，并内置了超过100个技能（Skills）以辅助完成任务。

相关链接：

• 昆仑万维发布“Skywork 桌面版”：AI 帮你在电脑上直接干活

Kunlun Tech Launches Desktop AI Assistant, Emphasizing Local File Processing Capabilities

Brief:

• On February 4th, Kunlun Tech globally released its desktop AI application, "Tiangong Skywork Desktop" (SkyworkDesktop). 🌍
• The application's core feature is its ability to perform tasks directly on the user's local computer. It can read, summarize, and organize massive amounts of local files (e.g., documents, images, videos) and generate new content, all without requiring any file uploads to the cloud.
• It is positioned as an "OS Assistant" and a "Windows version of Claude Cowork," aiming to integrate into desktop office scenarios and handle complex tasks across various file types and formats. 🖥️
• The application supports choosing between various large models like Claude and Gemini, and it includes over 100 built-in skills to assist in task completion. ✨

Related Link:

• Kunlun Tech Releases "Skywork Desktop": AI Helps You Work Directly on Your Computer

英特尔官宣造GPU，直指英伟达AI霸主地位

简报：

• 英特尔CEO陈立武在思科人工智能峰会上宣布，公司将开始生产GPU，正式进军由英伟达主导的AI算力市场。
• 为执行此战略，英特尔已组建工程团队，由前Arm高管凯沃尔克·凯奇奇安负责，并聘请了曾在高通任职的首席GPU架构师埃里克·德默斯。
• 新GPU将专注于人工智能模型的训练与推理，并利用先进封装技术来应对AI应用中日益严重的存储瓶颈问题。
• 此举旨在打破英伟达在AI加速器领域超过80%的市场垄断，并为英特尔18A工艺节点建立完整的代工与产品生态。

相关链接：

• 英特尔进军GPU市场挑战英伟达
• 英特尔CEO宣布AI算力新战略

Intel Officially Announces GPU Production, Directly Challenging NVIDIA's AI Dominance

Brief:

• Intel CEO Pat Gelsinger announced at the Cisco AI Summit that the company will begin producing GPUs, officially entering the AI computing market currently dominated by NVIDIA. 🚀
• To execute this strategy, Intel has formed an engineering team led by former Arm executive Kevork Kechichian, and has hired former Qualcomm chief GPU architect Erik Demers.
• The new GPUs will focus on training and inference for AI models, utilizing advanced packaging technologies to address the increasing memory bottleneck in AI applications. 💡
• This move aims to break NVIDIA's over 80% market monopoly in AI accelerators and establish a complete foundry and product ecosystem for Intel's 18A process node. 🎯

Related Links:

• Intel Enters GPU Market, Challenges NVIDIA
• Intel CEO Announces New AI Computing Strategy