Group 1 - Oracle will collaborate with OpenAI on computing power and develop multiple data centers across the United States, indicating a positive outlook for the global computing power industry [1][2] - Baidu has launched its self-developed video generation model MuseSteamer, while Google has introduced the new Veo3 video generation model globally, highlighting the intense competition in large model development [1][2] - Beijing has released a three-year action plan for "Artificial Intelligence + Pharmaceutical Health," aimed at accelerating the integration of AI in the pharmaceutical industry [2] Group 2 - The computer industry index fell by 1.28% last week, while the CSI 300 index rose by 1.54%, indicating a 2.82 percentage point underperformance of the computer sector [3] - As of the last trading day of the previous week, the price-to-earnings ratio of the computer industry stood at 51.7 times, with 106 out of 359 A-share component stocks rising and 253 falling [3] - The computer industry is expected to see dual improvements in performance and valuation due to the recovery of demand, with investment opportunities identified in various sectors including the Xinchuang sector and AI sector [3]

Group 1 - Huawei's Pangu team announced the open-source release of the Pangu 7B dense and 72B mixture of experts models, but faced allegations of plagiarism from Alibaba's Tongyi Qwen-2.5 14B model, with a high similarity score of 0.927 in attention parameter distribution [2][3] - Huawei's Noah's Ark Lab responded that the Pangu Pro MoE model was developed and trained on its Ascend hardware platform and not based on other vendors' models [2] - An article published on GitHub by a self-identified member of Huawei's Pangu team claimed that the team fabricated technological breakthroughs and used competitor models for training [3] Group 2 - Tencent responded to user complaints about the new "AI search" feature in WeChat, clarifying that it integrates public information without using user privacy data [4][5] - Baidu announced its largest search business overhaul in a decade, allowing for over 1,000 characters in search queries and integrating AI writing and image generation capabilities [6] Group 3 - The 2025 Global Digital Economy Conference revealed a list of the top 100 talents in the AI field, with a significant representation of Chinese individuals [7] - DeepSeek is reportedly ramping up overseas recruitment, aiming to attract talent for positions focused on artificial general intelligence (AGI) [9] Group 4 - ByteDance has produced over 1,000 robots in two and a half years, with a long-term goal of achieving embodied intelligence [10] - Zhipu AI released and open-sourced the GLM-4.1V-Thinking series, a multimodal model with 9 billion parameters, demonstrating superior performance in various benchmark tests [10] Group 5 - Yonyou Network Technology submitted an H-share listing application to the Hong Kong Stock Exchange, marking a significant step in its internationalization strategy [14] - Wisdom Eye was included in KPMG's inaugural "China Health Technology Top 50" list for its innovative applications in healthcare AI [14] Group 6 - Baidu officially open-sourced the Wenxin large model 4.5 series, which includes various models with different parameter configurations [15] - DingTalk launched over 100 free templates for the e-commerce industry, integrating AI functionalities for various business needs [16] Group 7 - Siemens and other EDA companies confirmed the lifting of U.S. export restrictions on chip design software to China, allowing for renewed access to their technologies [17][18] - Trump announced new tariffs set to take effect on August 1, with rates potentially reaching up to 70% [19] Group 8 - Microsoft is set to lay off nearly 9,000 employees as part of a restructuring plan aimed at optimizing processes and reducing management layers [20] - Elon Musk's xAI company completed a $10 billion funding round to further develop its AI solutions and data centers [20] Group 9 - Google announced the global availability of its latest AI video generation model, Veo3, which significantly enhances video production capabilities [21] - CoreWeave became the first AI cloud service provider to deploy NVIDIA's GB300 NVL72 system, boasting high AI performance [22] Group 10 - Cursor apologized for a pricing communication issue regarding its Pro Plan and offered refunds to affected users [23] - Cursor's developer Anysphere hired two former executives from Anthropic to strengthen its leadership team [25] Group 11 - Microsoft is incorporating AI usage into employee performance evaluations, reflecting its commitment to integrating AI tools into daily operations [26] - Apple is considering using AI technologies from Anthropic or OpenAI for its Siri assistant, indicating a potential shift in its AI strategy [27] Group 12 - Meta established a new department called the "Meta Superintelligence Lab," recruiting several prominent figures from the AI industry [28] - Multiple European companies urged the EU to pause the implementation of the upcoming AI Act, citing concerns over its impact on innovation [29] Group 13 - Figma submitted its IPO application, aiming to list on the NYSE, following a previous failed acquisition attempt by Adobe [31] - Remark completed a $16 million Series A funding round to expand its online retail guidance services [32] Group 14 - Zhiyu Technology went public in Hong Kong, raising approximately 320 million HKD for research and international market expansion [37] - Domestic GPU company Sunrise raised nearly 1 billion RMB in funding to support its high-performance GPU development [38]

Core Viewpoint - The commercialization of video generation has made significant progress, with revenues from marketing and promotion now balancing out with investments [1][24]. Group 1: AI Video Generation Trends - AI-generated ASMR and animal sports videos have gained immense popularity on short video platforms, attracting millions of likes and shares [2][3]. - The release of Google's video generation model Veo3 in May has been a game-changer, enabling high-quality AI videos with synchronized audio, thus transforming content creation [5][11]. - The rapid advancement of AI content creation tools has led to a surge in creators leveraging these technologies, with many accounts emerging on short video platforms [3][6]. Group 2: Market Dynamics and Competition - The competitive landscape is evolving, with various players like 可灵 and 即梦 making strides in the AI video generation space, alongside Google's efforts [10][14]. - 可灵's video generation model has achieved over 30% market share, surpassing competitors like Runway and Veo-2 [14]. - The industry is witnessing a shift in user preferences, with creators increasingly relying on video generation tools for efficiency, as evidenced by a threefold increase in download rates for generated images [15][19]. Group 3: Financial Performance and Projections - 可灵 is projected to achieve an annual recurring revenue (ARR) exceeding $100 million by March 2025, outpacing other AI products like Cursor [17][19]. - The annual revenue for leading video generation products is expected to reach $1 billion this year, with potential growth to $5-10 billion next year [19]. - Despite the positive outlook, industry leaders acknowledge that the commercialization process is still in its early stages, with many challenges remaining [25][26].

Core Insights - The advertising industry is increasingly adopting AI technologies, with major companies like TikTok and Meta launching new AI advertising tools to streamline the creation of video content [1][2][3] - The cost-effectiveness of AI-generated advertisements is a significant advantage, with some tools reducing production costs by up to 95% compared to traditional methods [3][5][6] - Despite the benefits, there are concerns about the quality and reception of AI-generated content, particularly regarding the "uncanny valley" effect and the potential for homogenized advertising [19][20][30] Group 1: AI Advertising Tools - TikTok introduced a new AI advertising feature that generates 5-second video ads from images or text prompts [1] - Meta upgraded its image-to-video advertising tool, allowing marketers to create multi-scene video ads using AI [2] - Google's Veo3 tool can create complete videos from a single prompt, significantly reducing production time and costs [3] Group 2: Cost Efficiency - AI-generated advertisements can be produced at a fraction of the cost of traditional ads, with some companies claiming costs as low as $1 per ad compared to $200 previously [6][8] - The use of AI tools allows for faster production timelines, with some ads being completed in just a few days [3][8] Group 3: Industry Trends - A significant portion of advertisers (53.1%) are already using AI-generated content in their marketing strategies, indicating a shift towards AI integration in advertising [2] - The trend towards AI in advertising is expected to continue, with many small and medium-sized companies adopting these technologies to remain competitive [11][13] Group 4: Quality Concerns - There are growing concerns about the quality of AI-generated content, with some users expressing dissatisfaction with the "creepy" appearance of AI-generated characters [19][20] - Reports indicate that AI-generated ads may lack the emotional connection and engagement that traditional ads provide, leading to a negative perception among consumers [23][24][30] Group 5: Future Outlook - The debate over the effectiveness of AI-generated content versus traditional advertising is likely to persist, as companies weigh the cost benefits against potential drawbacks in consumer engagement [31][33] - While AI tools are becoming more accessible, there remains a belief in the value of human creativity and expertise in advertising [18][33]

北京时间6月24日凌晨，微软在官网发布了创新小参数模型Mu。Mu仅有3.3亿参数，但其性能可以比肩 微软之前发布的小参数模型Phi-3.5-mini，体量却是其十分之一，并且在离线NPU的笔记本设备上，可 以跑出每秒超过100 tokens的响应，这在小参数模型领域非常罕见。此外，Mu支持在Windows中设置智 能体，可将自然语言指令实时转化为系统操作。 【观网财经丨智能早报 6月24日】 网易有道开源首个专注数学教育的模型 6月23日，智通财经获悉，网易有道宣布开源"子曰3"系列大模型的数学模型（英文名称Confucius3- Math）。这是国内首个专注于数学教育，可在单块消费级GPU上高效运行的开源推理模型。（智通财 经） 微软发布Mu模型：支持Windows智能体，小参数跑出10倍性能 抖音上线AI饭搭子"探饭"，知情人士：目前仅支持小程序 有媒体报道称，字节跳动的用户增长团队做了一个名为"探饭"的AI产品，搭载的是豆包大模型，其开发 公司为北京有竹居网络技术有限公司，经查询，为字节跳动的全资子公司。据悉，探饭主要为用户提供 美食相关的智能向导服务。 6月23日，据知情人士消息称，"探饭"是抖音生活 ...

昨天晚上刷X时，看到两个帖子，连起来看，非常有意思。 第一个帖子，是一个动物奥运会类型的AI视频爆了，原视频在Tiktok上都290万的点赞了，又在X和tiktok上掀起了一波潮流。 你在Tiktok上搜一下，又有一堆的模仿的人，拿到了很多不错的流量。 第二个帖子，是我的好朋友，汗青的一段提问。 连在一起看，非常的有意思，本来想在X上直接回复写一点点我的想法，但是思绪很乱，想说的话有很多，想着想着，我跟汗青说。 我不知道作为一个AI创作者和自媒体，对于流量的观点对大家有没有价值，但是我觉得还是可以表达一下。 当然，都是我的一家之言，不保证对。 正如汗青所说，最近AI视频类的爆款有点多。 先是6月初，用Veo3做的，一只举着自拍杆的卡皮巴拉，拍的Vlog爆火。 从而引发了一股Veo3的Vlog热潮，比如什么大脚怪自拍vlog、重生之我在三国做战地记者等等。 然后就是今天的这个，动物奥运会。 可能是最近Veo3的声音升级，还有hailuo 02的物理升级，确实解锁了很多以前做不了的玩法。 但是确实以AI驱动的爆款类型越来越多了。 回到汗青的那三个问题： 1. AI不断制造爆款视频格式的流行（近期已是事实）。那么从 ...

我的X、TikTok、抖音、小红书已经被一种视频霸屏了。 特别是小红书，没事就能刷到一个。 这种视频长这样，相信我，你一定要，带着耳机，放着声音听，要不然你会后悔的。 现在是凌晨1点，这是我听这个视频的第八遍，我不知道你们什么感觉，我自己还是有一种，皮肤触电、后脖发热，没事一抽抽的奇妙爽感。 这就是最近，火遍全网的AI ASMR。 原作者是TikTok上的一个新号，名字叫impossibleais。 6月11号才发的第一个视频，截止到现在，已经有17万粉丝了，涨粉涨的非常夸张。。。 再看看他的视频数据，我。。。我真的羡慕死好吧。 动不动就是几十万的点赞，甚至还有两个百万级的点赞。 羡慕的话真的说不动了。 给大家看一下，点赞量最高的，250万点赞的那个AI ASMR视频，记得一定要带耳机哦。 夜里听实在是听的浑身起鸡皮疙瘩。 再比如，扔一段黄金草莓的。 "Realistic 4K footage close up of a knife rapidly cutting a glowing gold glass strawberry on a wooden cutting board. It slices it m ...

「 一年内是吃模型红利头一口汤的黄金期，创业要趁早。 」 整理 ｜ 于丽丽 在中国的风险投资界，创新工场联合CEO兼管理合伙人汪华，素来以 颇具前 瞻性著称。2008年，他超前预测了移动互联网大潮的 到来，2011年则对移动互联网做出"三步走"的趋势判断。这些"预言"在后来的现实中，一一印证。 " 42章经"的曲凯， 因此曾认为，关于移动互联网，从对趋势的预测，和战略布局来说，只有两个人是天才级存在：一个是美团的 王兴，另一个就是汪华。 前不久，在杭州举办的 36氪WAVES新浪潮2025大会上，汪华发表了题为《超越连接，拥抱实现：Agent与多模态的十倍机会》的主 题演讲。 演讲中，他指出了AI时代与移动互联网时代范式的不同，并认为我们即将迎来一个遍地是黄金的AI时代,而所有AI创业者将分享有 生以来最大的一波红利：模型红利。 以下是演讲内容，经编辑： 非常高兴有机会和大家交流关于AI创业里的一些看法和我的观察。"超越连接，拥抱实现"，这也是过去两三年时间，和非常多的创 业者，包括投资人了解后的一些体会。 Part 01 移动互联网的旧范式 与 AI 时代的新范式 首先，我们讨论旧的范式和新的范式。AI时代 ...

前天MiniMax的M1文章里，我说MiniMax得掏一个视频模型出来吧。 于是，果然，前天深夜，他们发了Hailuo 02。 这个模型，前天深夜还不能用，还没正式上线，但是放了预告片。 他们的预告片，给我看的鸡皮疙瘩起来了，我给大家看一眼。 不是，这真的太离谱了，第一次，见到能生成杂技动作的AI视频模型，这个肢体表演和动作幅度，你们可以去Veo3之类的试试，是完全达不到的。 我转完朋友圈，朋友都惊了... 直接冲着所有视频模型的弱点开大。。。 所有人其实都知道，体操，被誉为AI视频的图灵时刻。 谁能解决体操的视频生成问题，谁就是真神。 而杂技的难度，也仅次于体操了。 所以昨天白天Hailuo 02上线了之后，我就直接做了一波新鲜的测试。 比如上周末，我二刷了《碟中谍8》。 虽然前面的文戏又臭又长让人犯困，但到了熟悉的动作戏段落，看到阿汤哥又是潜入深海又是高空扒飞机，我还是真情实感地喊出了牛逼。 一个六旬老人，兢兢业业跑了三十年，不顾安危，一直用血肉之躯挑战mission impossible，就为了让你在银幕上看到最炸裂的视觉效果，我想不出 比这更热血的事了。 那时候我在电影院里还在感慨，扒飞机这样的动作， ...

这仍然只是技术迭代的初期。 AI视频生成赛道又热闹起来了。4月，快手可灵发布2.0视频生成模型，6月，字节跳动发布了即梦3.0 Pro视频模型，就在昨天，谷歌宣布Veo3正式上线，今 天MiniMax也加入混战队列，开始卷性价比了。 6月18日，MiniMax在海内外官方平台宣布，新视频生成模型海螺AI（海螺02）上线，同时发布了一段大秀杂技的AI视频。官方表示，这段视频是由3位艺术 家耗时1.5天，使用海螺02生成的多个6-10s视频，再拼接剪辑而成。 杂技画面对AI视频生成来说一直是较有难度的内容，此前AI生成画面时往往肢体拼接混乱，无法准确模仿复杂的人类动作。从此次画面效果来看无论是光 影、人类动作、物理模仿都完成得很好。 不过，需要指出的是，有AI创作者对第一财经表示，AI视频生成还涉及一个成功率的问题，因此样片是完美的，但过程中他们进行了多少次"抽卡"，外界并 不能知晓。所谓抽卡指的是， AI往往不是一次就能成功生成用户想要的画面，可能会有出错的概率，因此行业会选择多次生成直到得到自己想要的画面。 不过，即便如此，上述AI创作者认为，这次海螺AI的更新是不错的，行业整体水平都在提升。 在竞技场Art ...