以下文章来源于CSDN ，作者万有引力 CSDN . 成就一亿技术人 作者 | 万有引力 出品 | CSDN（ID：CSDNnews） 当 Manus 以其惊艳的自主任务执行能力点燃 AI Agent 领域的热潮时，其"一码难求"的现 状也让众多开发者望而却步。几乎在同时，一个名为 OpenManus 的开源项目以"火箭 般"的速度问世，不仅成功复刻了核心功能，更以完全开放的姿态，在短短不到一个月的时 间内于 GitHub 吸引了超过四万 Star 数的关注（截止本文发布，项目 Star 数已经达到 42.2k）。 OpenManus 项目 Star 数 这一现象背后，站着一群充满活力的 00 后程序员。他们利用下班后的短短三小时，凭借对 技术的热爱与开源精神，迅速将一个想法变成了现实。这种惊人的执行力与纯粹的"Just for Fun"动机，引发了业界的广泛讨论：这一代年轻开发者是如何学习、成长并拥抱前沿技 术的？他们与 AI 工具的深度协作达到了何种程度？支撑他们快速行动的技术积累和开源理 念又是什么？OpenManus 的诞生仅仅是复刻吗？其技术内核与未来方向又将如何演进？ 梁新兵 ： 向劲宇 ： Op ...

2025 年 3 月全球 AI 重要趋势。 文 丨 贺乾明 2025 年 3 月的 AI 月报，我们开始尝试一种新形式：和知乎一起举办 "AI 脑暴" 线下活动，围绕每月 一个热门 AI 主题，邀请相关学界研究者、业界从业者（研发或技术人员）、投资人等一起做闭门圆 桌讨论。 3 月 30 日，第一期 AI 脑暴举行，主题是 Agent，我们邀请 6 位嘉宾参加，他们来自高校、互联网公 司和非营利 AI 研究机构。本期月报中，我们摘录了部分 AI 脑暴中的讨论。 本期月报，你会看到： 为什么开发 Agent 简单，但做好很难 AI 脑暴活动中 "壳重要还是模型重要" 部分讨论要点 以下是我们第 5 期 AI 月报，欢迎大家在留言区补充我们没有提到的重要趋势。 开发 Agent 简单，让它好用难 3 月初，通用 Agent 产品 Manus 上线第二天，就被复刻出两个开源版本——OpenManus、OWL。 OpenManus 甚至只用 4 个人，花 3 个小时。 这种 "速成" 似乎暗示：做 Agent 没那么难。但从实际体验和系统复杂度来看，让 Agent "真正好用" 仍存在挑战。 大模型公司 Anthro ...

更意想不到的是， 推动 OpenManus 诞生的核心开发者，竟是一群 00 后 ！ 这些年轻的程序员 完全出于兴趣和热爱 ，利用自己的时间写代码，探索 AI 的更多可能，只为让智 能体工具触手可及。 没有 KPI，没有商业利益驱动， 只有纯粹的技术信仰——Just for Fun 。 本月初， Manus 横空出世，迅速爆火！它凭借云端自主执行、多智能体协同、持续学习与记忆等核 心能力，无需过多的人工干预，就能直接交付完整的成果，也可以灵活调用各类工具，不仅能 写代 码、查资料、智能浏览网页 ，还能 操作各类应用 ，俨然一位"全能选手"。 因此，Manus 被不少人称之为"全球首个通用 AI 智能体"，瞬间点燃了 AI 圈众人的热情。无数开发 者看到 Manus 的惊艳 Demo 后迫不及待想要体验。然而， 邀请码成了最稀缺的"硬通货" ，一码难 求，让许多人望而兴叹。 也就在 此时，国 内专注于多智能体系统的技术公司 DeepWisdom 的 MetaGPT 团队迅速行动， 复 刻 Manus 并开源，推出了 OpenManus，直接把门槛打了下来—— 无需邀请码，所有人都能免费 用！ 同时，从复刻到上线 ...

Core Insights - DeepWisdom has completed a financing round of over 100 million yuan, with its product mgx.dev achieving a remarkable 1 million USD ARR in its first month without any promotional expenses [1] - The product has topped the Product Hunt global rankings for four consecutive weeks, showcasing its strong market reception [1] - The company is currently undergoing another round of financing, with Xinghan Capital serving as the exclusive financial advisor [1] Group 1: Product Development and Innovation - MGX (MetaGPT X) aims to solve the productization issues of MetaGPT, focusing on natural language programming [3][23] - The development of OpenManus was accomplished by a small team of four recent graduates in a remarkably short time, highlighting the effectiveness of their academic cycle [3][14] - The architecture of MGX consists of three layers: an operating system, an integrated development environment (IDE), and application production and distribution [24][25] Group 2: Academic Cycle and Team Dynamics - The concept of an academic cycle is crucial for innovation, allowing every team member to contribute incrementally to the organization's progress [10][12] - The success of organizations like OpenAI and ByteDance is attributed to their ability to foster an academic cycle, which encourages continuous contribution and improvement [10][12] - The team culture at DeepWisdom emphasizes transparency and autonomy, allowing members to take initiative and encouraging a trial-and-error approach to innovation [41][51] Group 3: Market Trends and Future Directions - The growth rate of MGX's ARR is claimed to be the fastest in Chinese history, indicating a strong market demand for its solutions [45] - The future of mobile interfaces may shift towards AI-generated applications, moving away from fixed applications [28] - The company aims to redefine programming paradigms by enabling users to generate code through natural language, positioning itself as a potential universal problem solver [23][29] Group 4: Challenges and Management Philosophy - The management philosophy at DeepWisdom is characterized by a lack of hierarchical titles, promoting a flat organizational structure where decisions are made collaboratively [48] - The company faces challenges in resource allocation and team bandwidth, necessitating ongoing optimization of its operations [55] - The emphasis on critical thinking and the absence of ego in discussions are key to fostering a productive team environment [51]

4 月 18-19 日，由 CSDN&Boolan 联合举办的 2025 全球机器学习技术大会（ML-Summit 2025）将在上海虹桥西郊庄园丽笙大酒店隆重举行。大会 云集院士、顶尖学者、IEEE Fellow、一线科技企业技术实战专家组成的超 50 位重磅嘉宾。他们将以独特的视角，解读智能体、联邦学习、多模态大 模型、强化学习等覆盖 AI 当下热门的技术实践专题。 在大会首日下午的「AI 智能体」专题论坛上，来自 DeepWisdom 算法研究 员、OpenManus 项目一作，MetaGPT 开源核心贡献者梁新兵将带来 《通用 Agent 的构建与赋能：OpenManus 的实践与探索》分享。 梁新兵是 DeepWisdom 算法研究员，华东师范大学硕士。他不仅是 OpenManus 项目一作，同时也是论文 Data Interpreter /Self-Supervised Prompt Optimization 作者之一。如今，他正以其在智能体领域的丰富经验和 对开源的满腔热情，不断探索通用 Agent 的构建与赋能。 3 小时复刻传奇：OpenManus 背后的极速行动 作为 MetaGPT 开源 ...

Core Viewpoint - The article discusses the rapid development and popularity of the intelligent agent sector, particularly highlighting the impact of the Manus product and the emergence of open-source projects like OWL and OpenManus, which have sparked a wave of innovation and competition in the field [1][2][3]. Group 1: Manus and Its Impact - Manus has significantly influenced the intelligent agent landscape, leading to a surge in both open-source and commercial closed-source products [1]. - The official social media account of Manus faced a temporary ban but has since resumed, promising more demonstrations and updates [12]. - Manus has gained traction internationally, with strategies such as distributing invitation codes to influencers and users [13][14]. Group 2: Open-Source Projects - The OWL project, developed by the CAMEL-AI team, has integrated the DeepSeek model into a multi-agent collaboration framework, showcasing its capabilities [3][4]. - OWL achieved an average score of 58.18 in the GAIA benchmark, ranking first among open-source projects [5][6]. - The CAMEL-AI team expressed confidence in improving their scores in the GAIA benchmark, despite some gaps in Level 2 and Level 3 scores compared to competitors [7]. Group 3: GAIA Benchmark - The GAIA benchmark, created by Meta AI, Hugging Face, and AutoGPT teams, consists of over 450 complex questions designed to evaluate the capabilities of intelligent agent systems [24][25]. - The benchmark is divided into three levels of difficulty, with Level 1 requiring simple problem-solving and Level 3 demanding advanced capabilities [26][27]. - Manus scored 57.7% in Level 3, significantly outperforming other systems, while its Level 2 score was close to that of commercial systems [28][29]. Group 4: User Experiences and Market Trends - Users have reported high satisfaction with Manus, noting its ability to accurately gather personal information and perform complex tasks [18][19][20]. - The willingness to pay for Manus is higher among international users compared to domestic ones, as it offers a more affordable alternative to other high-end AI solutions [17]. - The article highlights a growing interest in agent-related projects on platforms like GitHub, indicating a trend towards the development of specialized intelligent agents in various fields [8][9].