AI前沿每周关键词Top50 （ 0707-0711 ） 扫码 加入AGI数据库，AI智能问答 腾讯研究院 ： 《算法破茧｜腾讯研究院三万字报告》 | 类别 | Top关键词 | 主体 | | --- | --- | --- | | 模型 | Grok4 | 马斯克 | | 模型 | Phi-4新版本 | 微软 | | 模型 | 开放权重模型 | OpenAI | | 模型 | SmolLM3 | Hugging Face | | 模型 | Skywork-R1V 3.0 | 昆仑万维 | | 模型 | BlueLM-2.5-3B | vivo | | 模型 | DeepSeek-R1外挂 | 上海交大 | | 模型 | HumanOmniV2 | 阿里 | | 模型 | Skywork-Reward-V2 | 昆仑万维 | | 模型 | DeepSeek增强版 | 德国TNG公司 | | 模型 | Sekai数据集 | 上海AILab | | 应用 | AI浏览器Comet | Perplexity | | 应用 | MedGemma 27B | 谷歌 | | 应用 | 生肖企鹅AI共创 | 腾讯 | | ...

只见随着六边形的不断旋转，小球错落有致地从开口下落。 发布还不到一天，马斯克的Grok4就已经被网友们玩疯了。 比如有网友报告，Grok-4已经 成功通过了著名的六边形小球氛围编程测试 。 克雷西 发自 凹非寺 量子位 | 公众号 QbitAI 拿着显微镜捉虫的网友发现小球在返回中心位置时会穿墙，但作者表示这是故意为之。 | Plutus � @PlutusCosmos · 17小时 | | | | | --- | --- | --- | --- | | The balls penetrate the walls when the go back to the center. Is it intended? | | | | | O3 | U | ♡ 74 | 111 2.5万 | | Flavio Adamo � @flavioAd · 17小时 | | | | | yes | | | | | () 1 | 11 | C 59 | 111 2.5万 | | SoyTeslike � @soyteslike · 16小时 | | | | | damn, already screenshotted but it ...

值得注意的是，在发布会上同马斯克一起分享的两位xAI的核心创始成员，都是华人。这也并非他们首 次与马斯克同台。早在今年2月的发布会上，他们就已经以核心成员的身份亮相。 坐在中间位置的吴宇怀，是出生于杭州建德的"95后"，今年刚满30岁，却已在AI领域崭露头角。他在多 伦多大学攻读机器学习博士，师从"深度学习之父"杰弗里·辛顿（Geoffrey Hinton）。博士期间曾在 Google DeepMind和OpenAI实习，毕业后加入Google研究院，并在斯坦福大学从事博士后研究。2023年 xAI成立时，他作为联合创始人加入，成为团队中最年轻的核心成员之一。 坐在右侧的Jimmy Ba，不仅是xAI的创始成员，也是吴宇怀的导师之一。他同样毕业于多伦多大学，其 2014年与Diederik P. Kingma联合提出的Adam优化算法已成为深度学习领域的标准工具，被几乎所有主 流AI模型引用。2025年，他凭借这篇论文获得ICLR时间检验奖，该奖项仅颁发给发表十年后仍对行业 产生深远影响的论文，含金量极高。 0:00 21世纪经济报道记者 董静怡 上海报道 当地时间7月9日，埃隆·马斯克（Elon Musk） ...

Core Insights - Amazon Web Services (AWS) is set to launch an AI agent marketplace next week, with Anthropic confirmed as a partner [1][2] - The marketplace aims to provide a unified platform for startups to offer AI agent services directly to AWS customers, potentially driving significant growth for partners like Anthropic [1] Group 1: AWS AI Agent Marketplace - The AI agent marketplace will be officially unveiled on July 15 at the AWS Summit in New York [1] - The marketplace will allow enterprise users to browse, install, and retrieve various AI agents on demand [1] - This initiative is seen as a solution to the current fragmentation in the distribution of AI agents among enterprises [1] Group 2: Anthropic's Role and Growth Potential - Anthropic is reportedly receiving a new multi-billion dollar investment from Amazon and is focusing on the AI agent sector for future development [2] - The company not only develops its own AI agents but also empowers developers to create customized solutions through APIs [2] - Anthropic's annualized revenue surpassed $3 billion by the end of May, indicating strong growth potential [2] Group 3: Competitive Landscape - AWS is not the first tech giant to enter the AI agent marketplace; Google Cloud and Microsoft have also launched similar platforms [2] - The effectiveness of these marketplaces for small to medium-sized AI startups and specific customer needs remains to be validated [2] - AWS will take a small commission from agent installation revenues, but the potential value of customer acquisition and new revenue channels is significant [2]

5年9个月前，我告别硅谷。 那时，中美科技脱钩严冬初露锋芒：海外据点在硅谷立足不稳，中资基金大幅撤退，301法案精准阻 击。硅谷依然是硅谷，孵化器灯火通明，投资人奔走如织。而我，也试图在东西之间找寻自己的坐标。 去年冬天，在北京寒冷而炽热的一间办公室，我突然又"听见了"硅谷——不是新闻、不是公告，是一种 被时间打磨的召唤，也是重新寻找的答案。 离开与归来，我写在了《5年9个月，再回硅谷：起初纪年》里。今天的这篇，是我回到硅谷后的记录， 是一个创投观察者的再定位，也是一位曾经的参与者，对时代叙事变迁的个人备忘。 于是再回硅谷，见证落幕与告别，看到重生与希望，感悟创新与未来。 是以为纪。 一、旧时代的终章：脱钩与解体 无需否认，中美科技跨境的叙事早就结束了。 6月的下午，我在盛名一时的硅谷中关村创新中心门口看着落寞的广告牌。7年前，那里车水马龙，人声 鼎沸。大领导到来时，中国背景的政府，央企，风投，机构，创业公司们将门厅挤得水泄不通。平时的 孵化器也是灯火通明。 如今那里门可罗雀，停车场空无一人，只剩待租招牌在风中孤单伫立。 几公里之外，是当年深圳花大价钱投入的硅谷创新中心，哪怕地处如日中天的英伟达斜对面，它也难 ...

"I don't trust OpenAI and Sam Altman. I don't think we want to have the most powerful AI in the world controlled by someone who is not trustworthy."Elon Muskhttps://t.co/MhFBAi9WJr ...

Investment Rating - The report does not explicitly state an investment rating for the industry or specific companies involved Core Insights - CoreWeave has become the first cloud provider to commercially deploy the NVIDIA GB300 NVL72 platform, featuring a fully integrated system with significant performance improvements, achieving 1.1 ExaFLOPS for inference and 0.36 ExaFLOPS for training, representing a 50% performance uplift over the previous generation [2][12] - The acquisition of Core Scientific allows CoreWeave to control over 1.3 GW of power resources, expected to save approximately $500 million annually in operational costs and avoid $10 billion in future rental expenses, marking a strategic shift towards a vertically integrated infrastructure platform [5][14] - CoreWeave's partnerships with major clients like OpenAI and Google position it to become a leading vendor in the AI cloud infrastructure market, contingent on its ability to deliver compute commitments consistently [5][15] Summary by Sections Event Summary - In July 2025, CoreWeave announced its commercial deployment of the NVIDIA GB300 NVL72 platform, utilizing a fully integrated rack system with advanced components, achieving significant performance and efficiency improvements [2][12] Technical Architecture - The GB300 NVL72 architecture includes 72 Blackwell Ultra GPUs, Grace CPUs, and BlueField-3 DPUs, enabling high-speed communication and efficient power management through liquid cooling [3][17] Strategic Moves - The acquisition of Core Scientific for $9 billion enhances CoreWeave's control over data center resources, reducing reliance on third-party providers and lowering deployment costs, establishing a competitive advantage in the AI cloud sector [5][14] - The report highlights the increasing divergence in the Neocloud landscape, with CoreWeave's rapid deployment capabilities and integration of hardware and software setting it apart from traditional cloud service providers [6][17]

图片来源： Hard Fork Z highlights Casey Newton 是科技领域撰稿人，常聚焦 AI 行业动态与深度对话； Kevin Roose 作为《纽约时报》科技记者，追踪 AI 发展及相关争议；二者均是播客节 目《 Hard Fork 》的主持人。 Sam Altman 是 OpenAI 首席执行官； Brad Lightcap 负责 OpenAI 业务管理，助力 AI 技术落地与生态构建。本次访谈是 Hard Fork 直播的一个片段，该访谈首发于 2025 年 6 月 26 日 Hard Fork 频道。 轻松又诚意满满的对话开始 ——AI 大佬们的 " 轻松开麦 " 时刻 Casey Newton ： 节目大概进行到一半了，我想简单地问一下，你们现在感觉怎么样？我们在这边玩得挺开心的，哇哦。你是认真的吗？（笑）不太可能 吧。好吧，那挺好。 Sam Altman ： 他想回去了 . Casey Newton ： 别这样，来吧，来吧。是的，是的。你好， Sam 。你好， Brad 。 Sam Altman ： 见到你真好。 Casey Newton ： 我太喜欢了，今天我们是现场直播， ...

Group 1 - The overall A-share market showed a slight upward trend on July 11, with major indices mostly rising, particularly in sectors like brokerage, biotechnology, insurance, and heavy machinery [1] - The recently popular Cloud Computing 50 ETF (516630) saw a slight increase, with top-performing holdings including Dongfang Guoxin, Fanwei Network, Zhongke Xingtou, and others [1] - xAI released its latest AI model, Grok4, which achieved a comprehensive score of 73, surpassing OpenAI's o3-pro (71) and Google's Gemini 2.5 Pro (70), indicating it is currently the strongest AI model [1] Group 2 - The competitive landscape for large models is taking shape, with companies increasing investments to become significant players in the field [2] - xAI plans to launch a professional coding model in August, a multimodal intelligent agent in September, and a video generation model in October, aiming for comprehensive AI coverage within three months [2] - OpenAI is set to release GPT-5 in the summer, while Google is testing the next-generation Gemini 3.0 series models, indicating ongoing competition and demand for computing power [2]

Core Viewpoint - The AI sector is experiencing significant growth, as evidenced by the performance of the Sci-Tech Innovation AI ETF and the developments showcased at the Global AI Summit in Geneva [3][4][5]. Group 1: Market Performance - As of July 11, 2025, the Sci-Tech Innovation AI Index rose by 1.93%, with notable increases in constituent stocks such as Star Ring Technology (up 13.26%) and Cambricon (up 5.48%) [3]. - The Sci-Tech AI ETF (588790) increased by 1.78%, reaching a latest price of 0.57 yuan, and has seen a cumulative increase of 2.56% over the past three months, ranking 3rd among comparable funds [3]. - The latest scale of the Sci-Tech AI ETF reached 44.48 billion yuan, marking a new high since its inception and ranking 1st among comparable funds [4]. Group 2: Fund Flow and Investment Trends - The Sci-Tech AI ETF recorded a net inflow of 50.54 million yuan, with four out of the last five trading days showing net inflows totaling 118 million yuan [4]. - The latest financing buy-in amount for the Sci-Tech AI ETF was 13.30 million yuan, with a financing balance of 252 million yuan, indicating continued interest from leveraged funds [4]. Group 3: Historical Performance and Fees - The Sci-Tech AI ETF has seen a net value increase of 10.72% over the past six months, with a maximum single-month return of 15.59% since inception [5]. - The management fee for the Sci-Tech AI ETF is 0.50%, and the custody fee is 0.10%, which are relatively low compared to comparable funds [5]. - The tracking error for the Sci-Tech AI ETF over the past six months is 0.030%, the highest tracking precision among comparable funds [5]. Group 4: Index Composition - The Sci-Tech Innovation AI Index consists of 30 large-cap companies that provide foundational resources, technology, and application support for the AI sector [6]. - As of June 30, 2025, the top ten weighted stocks in the index accounted for 68.03% of the total index weight, with companies like Cambricon and Lanke Technology leading the list [7].