为什么 Agentic 叙事下 infra 是确定性最高的机会？过去 Agent 开发大多依赖开发者手动使用传统 Infra 搭建，开发工程量大、流程复杂，如何缩小 Agnet 的开 发难度和周期是一个确定性需求。随着 Agent 开发需求彻底爆发，更多人才和资源投入到新工具搭建中，Agent 开发范式的重构和收敛已经在发生。 另一个视角则是： 如果我们把 Agent 看做互联网使用者和开发者的增量，这些"新用户"的爆发式涌入又会带来什么样的需求？ 本文来自 Agent Infra 创业者 Hang Huang 的投稿，他以自己过去一年通过 AI 辅助进行产品开发的实践经验出发，为我们分享了第二个视角下 Agent Infra 的 在这篇文章中，他提了一个很有意思的观点" AI 的终极目标，不是帮人写代码，而是获得对整个软件生命周期的控制权 ：从构思到上线，再到持续运维"， 随着 AI 对工具利用深度和主动程度不断变化， AI-native Infra 也在相应进化，Hang Huang 把 AI-native 的 Infra 划分为了 5 个阶段，L0-L5，"在L5 阶段的 Agent-Native OS ...

周四A股商场高开高走、小幅震动上行，早盘股指高开后震动上行，盘中沪指在3368点邻近遭遇阻力，午 后股指维持震动，盘中软件开发、互联网服务、消费电子以及计算机设备等职业表现较好；珠宝首饰、贵 金属、美容护理以及食品饮料等职业表现较弱，沪指全天基本呈现小幅震动上扬的运行特征。当时上证综 指与创业板指数的均匀市盈率分别为13.69倍、35.80倍，处于近三年中位数均匀水平，适合中长时间布 局。两市周四成交金额12136亿元，处于近三年日均成交量中位数区域上方。北向资金接连净流入，南向 资金加仓港股；ETF份额回升，产业资本净增持，商场流动性韧性增强。估计短期商场以稳步震动上行为 主，主张以方针导向为主线，兼顾成绩确定性与估值修正潜力。仍需亲近重视方针面、资金面以及外盘的 变化状况。短线主张重视软件开发、互联网服务、计算机设备以及消费电子等职业的出资机会。 指数继续在合理水平震动。未来，商场的机会将取决于商场的增量性变化，商场如因外部风险、基本面、 出资者心情等因素导致意外下行，则从"安稳"的维度看，维稳资金的呵护将有助于商场阶段性底部的形 成，商场也将由此呈现下有底的特征。如果商场接连震动特征，则应结合高质量开展 ...

锐智开高：工业物联的隐形引擎 聚焦物联网与边缘计算，锐智开高在临港新片区的智慧港口建设中大显身手。其5G+AI集装箱调度系统 将故障响应速度提升50%；为长三角新能源汽车生产线部署的设备预测性维护平台，故障率降低40%。 技术纵深体现在毫秒级响应的边缘计算框架，某汽车零部件厂商的智能仓储系统通过其供应链协同平 台，实现库存周转率提升18%。 更值得关注的是其产学研协同模式——与上海交大共建"工业数字化联合实验室"，推动国产工业软件在 多协议兼容、碳中和监测等领域的技术突围。 技术生态构建者：多元路径赋能产业 1. 百度智能云：AI驱动的智能化引擎依托百度大脑的AI能力，在自然语言处理、图像识别领域表现突 出。其语音识别技术助力某汽车品牌APP操作效率提升50%，风控模型显著降低金融欺诈风险。云端一 体化开发工具正降低算法部署门槛，加速企业智能化进程。 在数字化转型的浪潮中，上海作为科技创新的前沿阵地，孕育了一批以技术深耕与场景落地见长的软件 开发力量。这些企业不再局限于单一技术输出，而是深入产业肌理，成为实体经济的数字化伙伴。 锐智互动：全场景数字化的架构师 从上海自贸区起步，锐智互动以"技术+场景"双轮驱动 ...

来源丨新智元（ID：AI_era） 图源丨Midjourney 印度老哥是真的有点猛啊！ 今天要说的这位，是 AI编程公司Builder.ai的创始人兼前CEO—— Sachin Dev Duggal。 是的，这家公司的后台里并没有AI，有的只是一群印度老哥假装AI写代码。 更劲爆的是，这一骗愣是坚持了8年。 但这周他算是彻底玩完了。 编辑丨 定慧 好困 他不仅造了个「全是人工，没有智能」的假AI公司 从软银、微软等巨头手里骗到了数亿美元融资，估值干到15亿 而且还敢对投资人虚报300%的营收 随着最近一次「欺诈」的曝光，上一轮的投资人吓得赶紧冻结了投资账户里的剩下的3700万美元（共 投资5000万美元），只给公司账户留下500万美元，而这500万美元还受限于政府的资金出境规定， 也没法用来发工资。 没办法，Builder.ai只能申请破产，此时的CEO早已换成来「擦屁股」的Manpreet Ratia——创始人 Sachin Dev Duggal于2月份辞去CEO职务，由Ratia接替。 这场闹剧直接导致了自2022年ChatGPT发布以来，AI初创公司中规模最大的一次倒闭事件——这家公 司在上一轮融资中 ...

Core Insights - Builder.ai, once valued at $1 billion, has declared bankruptcy, with debts exceeding $115 million owed to Microsoft and Amazon [1][7][9] - The company's business model was criticized for relying heavily on human labor rather than actual AI technology, leading to allegations of being a "fake AI platform" [3][5][11] - Despite raising nearly $500 million in funding, Builder.ai's operational costs far exceeded its revenue, with daily expenses estimated at over $500,000 [6][9][10] Company Overview - Builder.ai was founded in 2016 in the UK, initially named Engineer.ai, and launched its first AI platform product in 2017 [3] - The company claimed to offer a no-code app development platform that could reduce development costs to one-third of traditional methods and double the speed [3][4] - The platform's actual development was largely performed by low-cost engineers in India, contradicting its AI-driven claims [4][5] Financial Performance - Builder.ai raised significant funding, including $250 million in May 2023 and $50 million from debt markets in October 2024 [6] - The company reported revenue of £23 million for the fiscal year ending March 2023, covering less than 9% of its daily operational expenses [9][10] - The company faced severe cash flow issues, with only $7 million in cash reserves when the new CEO took over in March 2025 [10] Leadership and Management Issues - Sachin Dev Duggal resigned as CEO in March 2025, with Manpreet Ratia taking over amid ongoing financial struggles [10] - The company attempted to cut costs by laying off employees but ultimately failed to recover from its financial burdens [10][11] Industry Context - Builder.ai's downfall highlights the risks associated with the AI investment boom, where many companies may lack genuine technological foundations [11][12] - The situation serves as a cautionary tale for investors and developers, emphasizing the need for a return to practical, technology-driven solutions rather than speculative ventures [11][12]

IT之家 5 月 22 日消息，英国科技初创企业 Builder.ai宣布进入破产程序，该公司曾是英国融资最多的科技初创企业之一。 Builder.ai总部位于伦敦，是一家专注于利用人工智能技术简化应用程序和网站开发流程的公司。该公司曾宣称其技术能够将开发过程变得"像点披萨一样简 单"，并获得了包括微软和卡塔尔主权财富基金在内的多家顶级投资者的支持，累计融资超过 5 亿美元（IT之家注：现汇率约合 36.01 亿元人民币）。 Builder.ai在周二的一次公司全员电话会议上告知员工，公司正在申请破产。该公司向英国《金融时报》证实，其主要子公司 Engineer.aiCorporation"将启动破 产程序，并将指定一名管理人来管理公司事务"。 今年早些时候，公司创始人 Sachin Dev Duggal 辞去了首席执行官职务，但仍保留了董事会席位，并担任"首席巫师"（Chief Wizard）一职。4 月，公司新任 首席执行官 Manpreet Ratia 表示，公司已将 2023 年的收入下调至 1.4 亿美元，并将 2024 年下半年的收入预测下调了 25%。 据内部人士透露，Ratia 在周二的全体员 ...

Group 1 - The core viewpoint of the articles highlights the strong performance of AI applications and the semiconductor industry, with the AI ETF (588760) gaining attention due to its large scale and liquidity [1][2] - Nvidia's CEO Jensen Huang stated that DeepSeek will increase AI computing demand by 100 to 1000 times, predicting that China's AI market will reach $50 billion by 2026 [1] - The AI ETF closely tracks the Shanghai Stock Exchange's AI index, which includes 30 large-cap companies providing resources, technology, and application support for the AI industry, with the top three sectors being semiconductors (53.1%), IT services (16.4%), and software development (10.3%) [1] Group 2 - As of May 21, the AI ETF (588760) has over 2.5 billion shares and a total size of 1.43 billion yuan, with over 2 billion yuan bought in the last four months [2] - The ETF has shown a daily tracking error of 0.0126% since March, ranking first among its peers in tracking error control [2] - Industry analysts believe that the continuous improvement of domestic large models is narrowing the gap with top global models, and 2025 is expected to be a breakout year for AI applications [2]

转自InFOQ 作者｜ Sebastian 编译｜傅宇琪 策划｜冬梅 早在 2002 年，Rod Johnson 就提出了对 Java 企业级开发的批判性看法，并推出了一种更加简洁、 灵活的替代方案——Spring 框架。20 多年后，这位从事编程 30 多年的元老级码农又燃起了对 Kotlin 的兴趣。那么，是什么让他弃 Java 转向 Spring？他又是如何看待 Kotlin 的未来的呢？ 最近，Rod Johnson 在播客节目中，与主持人 Sebastian 和 Márton 详细回忆了 Spring 诞生故事， 并分享了他最近重返 JVM 的历程，以及作为 Kotlin 新手的种种乐趣。基于该播客视频，InfoQ 进行 了部分增删。 核心观点如下： Spring 的诞生 Márton：Spring 是如何诞生的？ Rod： 我认为，许多成功的软件往往都源自于开发者曾经亲身经历过痛苦。因此，Spring 之所以能 够成功，其中一个原因是，最初参与 Spring 开发的许多开发者，很多是从企业应用开发领域走出来 的。 Spring 的诞生可以说有些特殊。许多核心理念，比如后来被称为"依赖注入"，我是在 ...

Core Viewpoint - Railtown AI Technologies Inc. has launched a Visual Studio Extension that integrates its AI-driven Root Cause Agent (RCA++) to enhance developer productivity and streamline workflows [1][2][3]. Product Features - The Visual Studio Extension offers automated fixes, intelligent code suggestions, and streamlined workflows tailored to individual developers' codebases [2][4]. - Key features include AI-powered recommendations, pull request generation, bulletproof unit testing, environment tracing, and integration with Azure DevOps and GitHub [6]. Market Strategy - Railtown AI has engaged i2i Media for a marketing awareness campaign to expand reach and drive adoption of its AI-powered development tools [5][7]. - The extension is designed for both individual developers and larger organizations, providing scalable deployment and enhanced team collaboration [3][4]. Company Overview - Railtown AI Technologies Inc. is based in Vancouver, Canada, and focuses on AI-powered developer productivity tools that aim to reduce downtime and accelerate innovation [8].

Group 1: Vertical Field Scene Deconstruction - The company Ruizhi Interactive has developed a smart classroom solution that has been implemented in 37 key high schools across the country, achieving a teacher-student interaction delay of 200ms, which is compressed to a non-perceptible range during teaching [1] - The latest HarmonyOS multi-terminal adaptation framework allows the same code to maintain pixel-level consistency across electronic class signs, teacher pads, and student terminals [1] - The retail team at Ruizhi has completed stress testing for a fresh supply chain app, which utilizes a self-developed inventory dynamic forecasting algorithm to automatically trigger procurement alerts based on price fluctuations in the Xinfadi wholesale market, resulting in a 21% year-on-year reduction in loss rates for a chain supermarket [1] Group 2: Infrastructure Technology Empowerment - Baidu's voice interaction lab has developed a barrier-free system integrated into a government service app that accurately recognizes regional sign language expressions, enhancing customer service response speed in financial apps by 3.8 times [4] - JD's digital twin supply chain system is optimizing the delivery paths of surgical instruments through UWB centimeter-level positioning technology, saving transportation mileage equivalent to running around the Forbidden City 8,920 times annually [4] - Tencent's social viral toolkit is transforming offline commerce, with a membership app for a trendy store in Sanlitun achieving a weekend conversion rate of 47% through a combination of LBS and AR fitting rooms [4] Group 3: Technological Anthropology in Hutongs - A micro-studio focused on cultural heritage digitization has developed an AR guide app for the Shijia Hutong Museum, which recognizes brick carving patterns through smartphone cameras and automatically plays corresponding oral histories [5] - An arts forum in the 798 Art District led to the creation of a literary app that adjusts message push frequency based on user reading duration, potentially being the first content product in the country to incorporate a reader's breathing rhythm algorithm [5]