通用人工智能

Search documents
国际产业新闻早知道:中美马德里经贸会谈成果公布,中国国产AI芯片抢夺市场窗口期
Chan Ye Xin Xi Wang· 2025-09-16 05:34
Group 1 - The core point of the China-US economic talks is the mutual benefit and win-win nature of the economic relationship, with both sides agreeing to resolve issues through cooperation and dialogue [1][2][3] - The talks resulted in a basic framework consensus on addressing TikTok-related issues, reducing investment barriers, and promoting economic cooperation [1][2] - Both parties emphasized the importance of stable China-US economic relations for global economic stability and development, committing to further discussions to enhance cooperation [2][3] Group 2 - The Chinese government is determined to protect its legitimate rights and interests, particularly regarding overseas Chinese enterprises, and will conduct technology export approvals in accordance with the law [1][3] - The talks are part of a series of discussions held since May, indicating a continuous effort to address economic issues through dialogue [2] - The Chinese side expressed hope for the US to take concrete actions to remove restrictions on China, fostering a positive atmosphere for the stability of China-US economic relations [1][2] Group 3 - The essence of the China-US economic relationship is highlighted as mutually beneficial, with both sides having vast cooperation space and common interests [1][3] - The Chinese side opposes the politicization and weaponization of technology and economic issues by the US, advocating for equal negotiations [3] - The need for the US to show sincerity and abandon unilateral pressure tactics is emphasized for achieving more results in the talks [3]
杭州城投携手智谱发布多个大模型 推动城市治理现代化
Zheng Quan Ri Bao· 2025-09-15 16:25
此外,全国首个路桥养护多模态大模型依托智慧巡检系统,实时采集路面图像与地理信息,精准识别病 害,甚至察觉绿化枯萎。全国首个自主学习清运调度智能体则实现清运点自动匹配、垃圾量合理预测及 不安全行为实时识别等功能,整体识别准确率超90%。 这些成果源于智谱与杭州城投共同构建的"AI城投大脑"平台,该平台以GLM-4.5模型和MaaS开放平台 为技术底座。正如智谱CEO张鹏所言:"AI的未来,不仅是模型性能的较量,更是生产范式的重构。在 这个时代,每一位开发者、设计者、创业者都是智能生态的共同缔造者。" 9月15日,杭州市城市建设投资集团有限公司(以下简称"杭州城投")与北京智谱华章科技股份有限公 司(以下简称"智谱")联合宣布,双方共同打造的杭州城投人工智能产业大模型项目(一期)建设成果 正式发布。 《证券日报》记者了解到,本次发布涉及多个"全国首个"大模型及智能体,不仅展现了AI技术与城市公 共服务的深度融合,更是对未来城市治理模式的重要探索。 具体来看,全国首个公交大模型致力于解决公交行业普遍存在的"下客"客流识别难题,准确率达到 93%,到站预测精准至分钟级别。全国首个AutoGLM防汛智能体可提前24小时 ...
罗永浩提议与贾国龙公开直播对质;宇树入选MIT“聪明公司”
2 1 Shi Ji Jing Ji Bao Dao· 2025-09-15 02:53
(原标题:罗永浩提议与贾国龙公开直播对质;宇树入选MIT"聪明公司") 21世纪经济报道新质生产力研究院综合报道 早上好,新的一天又开始了。在过去的24小时内,科技行业发生了哪些有意思的事情?来跟21tech一起看看吧。 【巨头风向标】 罗永浩提议与贾国龙公开直播对质 9月14日,西贝创始人贾国龙在某个行业群内的表态截图流出。贾国龙表示:"我应对方式有错,改。做饭的围着吃饭的转,你说咋好就咋办。"并 称"罗永浩是网络黑嘴,是网络黑社会,太坏了。但他打醒了我,算变相的帮西贝进步。"9月15日凌晨消息,罗永浩针对西贝创始人贾国龙在行业 群里发言并提及自身一事发文称:"贾总,你说我是网络黑社会,我认为你是诬蔑诽谤。这次的事件,总是我说几句,你说几句,容易各说各话, 媒体转来转去也容易出现信息偏差,我们还是找一个大的网络平台直播,当面公平公正冷静理性地对一次话吧。相信这也能澄清西贝的真相,并 且对中国预制菜产业和餐饮行业的健康发展做一些贡献@西贝贾国龙。" DeepSeek、宇树科技等被MIT科技评论评为聪明公司 9月12日,《麻省理工科技评论》"50家聪明公司"最新评选结果揭晓,DeepSeek、宇树科技等明星创企 ...
OpenAI据传大砍微软(MSFT.US)商业分成
智通财经网· 2025-09-14 01:05
Core Viewpoint - OpenAI and Microsoft have reached a non-binding memorandum of understanding regarding the next phase of their collaboration, indicating ongoing negotiations on contract terms [1][3] Group 1: Revenue Sharing and Financial Implications - OpenAI has informed some shareholders that the revenue share given to Microsoft will significantly decrease in the coming years, from an initial 20% until 2030 to an expected 8% by the end of the century, potentially allowing OpenAI to retain over $50 billion in additional revenue [3] - It remains unclear whether the non-binding agreement includes modifications to the revenue-sharing plan, with some OpenAI executives seeking to exclude certain unreleased products from the revenue-sharing agreement [3] Group 2: Cloud Services and Competitive Dynamics - OpenAI is currently negotiating the scale of its spending on Microsoft cloud services, with previous estimates suggesting a spending of $135 billion by 2030, while Oracle has reportedly secured a $300 billion order from OpenAI [4] - Microsoft shareholders are questioning why OpenAI is placing larger orders with Oracle compared to Microsoft Azure, despite Microsoft being a key stakeholder in OpenAI [5] Group 3: Ownership Structure and Governance - The terms regarding equity distribution have reportedly been largely determined, with OpenAI's non-profit parent company and Microsoft each expected to hold about one-third of the new company's shares, although Microsoft will not receive a board seat [5]
OpenAI据传大砍微软商业分成
财联社· 2025-09-13 23:41
种种迹象显示,正忙着为购买算力筹措资金的OpenAI,将目光瞄向本应该分给合作伙伴微软的钱。 北京时间周五早晨,OpenAI与微软发布了一条语焉不详的联合声明,宣布 双方就"下一阶段合作"达成了"一项非约束性谅解备忘录",并且"正 在积极推进合同条款的敲定"。 (来源:公司官网) 要知道,围绕着OpenAI将经营实体转为公共利益公司的事宜,双方已经谈了至少半年。赶在本周发布一条"积极推动条款敲定"的公告,离 不开甲骨文的影响,也说明背后的利益纠葛尚未谈拢。 根据最新爆料, 在与微软敲定新合作条款的同时,OpenAI已告诉部分股东,未来几年其交给微软的营收比例将大幅下降。 按照双方最初的合作协议, 微软有权获得OpenAI到2030年为止的20%收入。 而根据OpenAI的最新预期, 到本世纪末将与商业合作伙伴(主要 是微软)分享约8%的收入。 按照目前 的 营收测算,这两个比例之间的差额意味 着 OpenAI将额外保留超过500亿美元的收入。 目前尚不清楚微软与OpenAI宣布的非约束性协议中,是否包含收入分成计划的修改。据悉,OpenAI的部分高管也希望将部分尚未发布的产 品排除在分成协议外,例如每月费用可 ...
通知 | 关于征集“中国机电一体化技术应用协会具身智能分会”发起单位的通知
机器人圈· 2025-09-12 10:05
关于征集"中国机电一体化技术应用协会具身智能分会"发起单位的通知 中机电协〔2025〕48号 各有关单位: 为贯彻落实党的二十大提出的前瞻谋划未来产业,抢占科技革命制高点重要精神,本着充分发挥行业协会 的组织作用,整合资源协同解决具身智能在发展中遇到的挑战和问题,推动我国具身智能实现跨越式发展,为 经济高质量发展锻造新质生产力引擎。经研究决定,拟成立"中国机电一体化技术应用协会具身智能分会"。现 将有关事项通知如下: 一、具身智能分会成立背景 (一)具身智能战略价值与产业情况 2025年3月,中央政府工作报告明确提出培育具身智能未来产业。作为人工智能的前沿领域,具身智能融 合了人工智能、机器人技术、认知科学等多学科成果,使智能体能够在环境中完成感知、决策与行动的闭环, 实现高效人机协同,已成为新一轮科技革命和产业变革的关键驱动力。 当前,具身智能产业迅猛发展,仅2025上半年国内融资额突破230亿元。国际科技巨头也纷纷布局,具 身智能作为实现通用人工智能重要路径,在工业、物流、医疗、养老等领域潜力巨大。我国已涌现一批具身智 能领域创新企业,关键技术取得进展,场景应用持续拓展,多地出台政策持续加持。 (二)机电 ...
马斯克AI预言:明年超越个人,5年超越全人类
Sou Hu Cai Jing· 2025-09-12 07:11
近日,特斯拉 CEO 马斯克做客播客《All-In》,再次提出轰动性言论。马斯克表示,在短短五年内,人工智能的智慧总量将与全人类持平甚至超越。 在播客中,马斯克为xAI的发展制定了清晰且激进的时间表。他透露,团队计划在明年推出功能更为强大的Grok聊天机器人,并正式迈向 AI Agent 的新阶 段。这意味着AI将不再仅仅是被动回答问题的工具,而是能够理解复杂指令、自主规划并执行多步骤任务的智能实体。 然而,这仅仅是序幕。马斯克进一步预测,到2026年,人工智能在所有关键智能指标上都将超越任何一个单一人类。到2030年,也就是大约五年后,人工智 能的智慧将匹敌甚至超越全球所有人类智慧的总和。 "我们将见证一个远超我们想象的智能爆炸,"马斯克在节目中表示,"这个速度是指数级的,而大多数人仍在线性思考。" 这一观点无疑极具前瞻性,但也 引发了业界对其可能过于冒进的担忧。 为了支撑其宏伟蓝图,马斯克也谈到了xAI目前的技术进展。他承认,团队仍在全力优化Grok 4模型。据介绍,Grok 4的核心优势在于其原则推理能力,旨 在让AI的回答更具逻辑性、准确性,并大幅减少AI领域普遍存在的幻觉——也就是模型会捏造并输出看 ...
OpenAI与微软就合作关系修订达成初步协议,为重组扫清道路
美股IPO· 2025-09-12 01:38
OpenAI与微软宣布就修订合作关系达成初步协议,为重组成传统营利性公司扫清了关键障碍。根据重 组计划,重组后的非营利部门将继续控制新的公益公司,并拥有价值超过1000亿美元的股权,约占该 公司寻求的5000亿美元估值的20%。微软美股盘后一度涨2%。 OpenAI与微软宣布就修订合作关系达成初步协议,为重组成传统营利性公司扫清了关键障碍。 布雷特·泰勒强调: OpenAI作为非营利组织起步,今天仍然是,未来也将继续是。主导我们未来的,始终是非营利部分。 这一初步协议的达成,使得OpenAI首席执行官Sam Altman向其公司结构改革的目标迈出了重要一 步。Altman及其投资者认为,将OpenAI转变为更传统的营利性实体,对于公司在未来几年筹集所需 的海量资金至关重要。 协议宣布后,微软在盘后交易中一度上涨超2%。但该交易仍需监管部门批准,且面临包括马斯克在 内的多方反对。 解除旧有束缚 OpenAI的崛起一直伴随着其复杂的、非传统的公司结构带来的挑战。 该公司最初是一家非营利研究机构,后来设立了一个有利润上限的子公司,以便为微软等外部投资者 打开大门。 周四,OpenAI主席布雷特·泰勒(Bret Ta ...
北京计划到2026年底发展通用人工智能产业创新伙伴成员规模至1000家
Xin Hua She· 2025-09-11 14:51
新华社北京9月11日电(记者阳娜)在9月10日举行的2025年中国国际服务贸易交易会"北京人工智能生 态伙伴大会"上,《北京市通用人工智能产业创新伙伴计划2.0》(以下简称"伙伴计划2.0")正式发布。 伙伴计划2.0明确,北京计划到2026年底,通过综合评估、动态进出机制,发展通用人工智能产业创新 伙伴成员规模至1000家。 据介绍,北京市通用人工智能产业创新伙伴计划自2023年5月启动以来,已汇集算力、数据、模型、应 用、投资等五类超280家伙伴企业,梳理32个分领域场景需求,为伙伴企业提供约8500P算力支撑,与 100余家创投机构合作;累计形成近200个大模型行业应用典型案例,涵盖金融、政务、文旅、医疗、营 销等10余个领域,促成40余个项目合作签约。 伙伴计划2.0将在汇聚五类优质伙伴的基础上,新增构建"1+10+X"服务体系,聚焦产业供需对接这一核 心任务,实施芯模适配等10项能力工程,以及伙伴计划能力推广X矩阵。同时,伙伴计划2.0还重点创新 了"城市行"和"AI产品体验计划",面向伙伴计划城市行的国内城市、全球数字经济城市联盟的海外城 市,以及京津冀区域城市提供一定数量的AI免费体验产品及服务 ...
π0.5开源前,国内也开源了一个强大的端到端统一基础模型!具备强泛化和长程操作
具身智能之心· 2025-09-11 02:07
Core Viewpoint - The article discusses the release of π0.5 and WALL-OSS, highlighting their advancements in embodied intelligence and the significance of these models in the robotics industry, particularly in enhancing task execution in complex environments [1][3][5]. Group 1: Model Capabilities - π0.5 demonstrates enhanced generalization capabilities through heterogeneous task collaborative training, enabling robots to perform long-term, fine-grained operations in new household environments [3][5]. - WALL-OSS achieves embodied perception through large-scale multimodal pre-training, allowing seamless integration of instruction reasoning, sub-goal decomposition, and fine-grained action synthesis within a single differentiable framework [8][18]. - The model exhibits high success rates in complex long-term manipulation tasks, showcasing robust instruction-following abilities and understanding of complex scenarios, surpassing existing baseline models [8][18][28]. Group 2: Training and Data - The training process for WALL-OSS involves discrete, continuous, and joint phases, requiring only RTX 4090-level computational power for training and inference deployment [14][15]. - A multi-source dataset centered on embodied tasks was constructed, addressing the lack of large-scale, aligned VLA supervision and current visual language models' spatial understanding gaps [20][22]. - The dataset includes thousands of hours of data, focusing on both short-range operation tasks and long-range reasoning tasks, ensuring comprehensive training for the model [20][22][24]. Group 3: Experimental Analysis - Experimental analysis on embodied visual question answering and six robotic operation tasks focused on language instruction understanding, reasoning, and generalization, as well as planning and execution of long-term, multi-stage tasks [25][31]. - WALL-OSS significantly outperformed its original baseline model in object grounding, scene captioning, and action planning tasks, demonstrating its enhanced scene understanding capabilities [27][28]. - The model's ability to follow novel instructions without task-specific fine-tuning was validated, achieving 85% average task progress on known object instructions and 61% on novel object instructions [29][31]. Group 4: Industry Impact - The advancements in WALL-OSS and π0.5 are positioned to address existing limitations in visual language models and embodied understanding, paving the way for more capable and versatile robotic systems [5][8][20]. - The company, established in December 2023, focuses on developing a general embodied intelligence model using real-world data, aiming to create robots with fine operational capabilities [39]. - The recent completion of a nearly 1 billion yuan A+ round of financing indicates strong investor confidence in the company's direction and potential impact on the industry [39].