Core Viewpoint - The announcement of advertising testing on ChatGPT by OpenAI raises concerns about user privacy and potential manipulation, as users have shared sensitive information during interactions with the AI [1][2][8]. Group 1: Concerns about Advertising and User Data - OpenAI's strategy of incorporating advertising may lead to the exploitation of user data, which includes private thoughts and feelings shared during interactions with ChatGPT [2][4]. - The creation of a "human honesty archive" through user interactions poses risks if this data is used for targeted advertising, as there are currently no effective tools to understand or mitigate these risks [2][4]. - The potential for user manipulation increases if advertising is based on sensitive data, which could undermine the trust users have in the AI as a non-profit-driven entity [2][4]. Group 2: Alternatives to Advertising Models - The industry faces a dilemma between limiting access to transformative technology to those who can afford it or adopting an advertising model that exploits user vulnerabilities [3][11]. - A third option proposed is to maintain broad access to AI tools while restricting corporate surveillance and manipulation of users [3][11]. - OpenAI has stated that it will adhere to principles for advertising, ensuring clarity and non-intrusiveness, but there are concerns about future deviations from these principles as the company develops its economic engine [3][11]. Group 3: Governance and Oversight Mechanisms - The establishment of a governance structure that includes independent oversight of personal data usage is essential to prevent exploitation [6][13]. - Suggestions include implementing cross-subsidy mechanisms where profits from high-value commercial clients subsidize public access to AI tools [5][13]. - Another proposal is to manage user data through independent trusts or cooperatives, ensuring that user interests are prioritized [14]. Group 4: Potential Outcomes and Industry Implications - The trend of maximizing user engagement at the expense of ethical principles may already be occurring within OpenAI, with reports indicating a focus on daily active user metrics [4][12]. - The advertising revenue model could ensure that powerful AI tools remain accessible to a broader audience rather than just those who can afford high subscription fees [4][12]. - The industry must find a balance to avoid creating a system that either manipulates users or restricts access to technology based on financial capability [7][14].

1.【字节跳动：Seedance 2.0还远不完美，将持续探索大模型与人类反馈的深度对齐】2月12日，字 节视频创作模型Seedance 2.0正式发布。字节跳动官微发文表示，Seedance 2.0还远不完美，其生 成结果仍存在诸多瑕疵。未来将持续探索大模型与人类反馈的深度对齐，让更高效、更稳定、更具想 象力的音视频生产工具，服务更多创作者。据了解，该模型支持多模态全能参考，允许组合输入不同 文本、图片、视频、音频，同时新增视频编辑能力，支持对指定片段、角色、动作或剧情进行定向修 改，还提供了视频延长功能。目前，该模型已上线即梦AI、豆包等平台。（界面新闻） 欢迎加入 睿兽分析会员 ，解锁 AI、汽车、智能制造 等相关 行业日报、图谱和报告 等。 海联 17万+ t 2万 投资价值企业 O 投资机构 及上市公司 5077 国家高新 小时一手资讯 技术企业 C = = = = = 线 -20 FF 创投人物 投融资及收并购事件 热门产业 创新机会 T= l 谱 2万+LP数据 全牛命周期 10万+基金数据 独角兽企业 各维度权威榜单 1万+专精特新小巨人 1400 FITI ER 十子标签 产业图谱 行业标签 ...

Core Viewpoint - Anthropic has successfully raised $30 billion in its latest funding round, bringing its post-money valuation to $380 billion, significantly enhancing its competitive position against OpenAI [1][3]. Group 1: Funding and Valuation - The latest funding round was led by Singapore's sovereign wealth fund GIC and Coatue Management, with participation from notable investors including D.E. Shaw & Co., Dragoneer Investment Group, Founders Fund, Iconiq, and MGX [3]. - Anthropic's valuation has nearly doubled compared to previous rounds, making it one of the most valuable private companies globally. Just months prior, the company raised $13 billion, while OpenAI is planning to raise up to $100 billion [3][4]. - The company confirmed that employees will be allowed to sell shares at the same valuation as this funding round [4]. Group 2: Business Growth and Revenue - Anthropic, founded in 2021, focuses on developing safe and responsible technology, primarily targeting enterprise sales in high-margin sectors such as software engineering, finance, and healthcare [4]. - The company's annualized revenue has surged, increasing from over $9 billion last year to $14 billion recently, reflecting significant demand from clients [4]. Group 3: Technology and Market Impact - Anthropic's technology includes coding agents capable of writing and debugging software with minimal human intervention, which has already impacted global markets. A tool for automating legal work recently caused stock declines in related companies [4]. - The company has launched an AI model optimized for automating enterprise tasks, including financial research, which has also led to stock price drops in several financial services firms [4]. Group 4: Infrastructure Investment - Despite rapid business growth, Anthropic's investments in AI development and large-scale application infrastructure are substantial, with plans to invest $50 billion in building data centers in the U.S. and procure AI chips worth hundreds of billions from Alphabet [5]. - However, these investments are still modest compared to OpenAI, which has committed over $1.4 trillion for AI infrastructure in the coming years [5]. - Both Anthropic and OpenAI rely on funding support from large chip manufacturers and cloud service providers, raising concerns about potential "circular transactions" sustaining the entire industry [5]. Group 5: Strategic Partnerships - Microsoft and Nvidia have previously indicated a combined investment of up to $15 billion in Anthropic, with part of this amount already included in the latest funding round [6]. - Microsoft is a major supporter of OpenAI, while Nvidia is expected to invest $20 billion in OpenAI's upcoming funding round [6].

Core Viewpoint - The launch of the new flagship model GLM-5 by Zhipu has significantly impacted its stock price and the broader AI sector, indicating a strong demand for AI capabilities and advancements in the domestic AI market [1][2]. Group 1: Zhipu's Performance and Model Launch - Zhipu's stock price surged over 40% to reach a historical high of 402 HKD per share, closing at 28.68% higher at 402 HKD, with a market capitalization of 179.2 billion HKD [2]. - The GLM-5 model has shown exceptional performance in programming and agent capabilities, comparable to Claude Opus 4.5, and is optimized for complex systems engineering and long-range agent tasks [1]. - Zhipu has completed deep reasoning adaptation and operator-level optimization with major domestic chip platforms, enabling stable operation on domestic computing clusters [2]. Group 2: Market Trends and AI Sector Activity - The AI application concept sector in the A-share market saw an overall increase of 0.13%, with stocks like Kunlun Wanwei and Dongfang Guoxin rising over 10% [1][4]. - The software sector in the Hong Kong stock market also experienced a rise of 0.56%, influenced by Zhipu's stock performance [2]. - The recent surge in AI model launches in China has provided users with more options, with significant advancements in various AI applications, including video generation and programming [7][8]. Group 3: Pricing Adjustments and Demand - Zhipu announced a price adjustment for its GLM Coding Plan, increasing subscription prices by 30% to 60% and API call prices by 67% to 100%, reflecting strong market demand [3][2]. - The adjustments aim to enhance service quality and stability under high load conditions, with existing subscribers' prices remaining unchanged [2]. Group 4: Industry Developments and Future Outlook - The AI industry is witnessing a collective surge, with significant investments in AI-related capital and infrastructure, particularly in high-performance chips and cooling technologies [5]. - The demand for AI applications is expected to grow, driven by increased user engagement and advancements in AI capabilities across various sectors [8].

在以每秒输出100个token的情况下，连续工作一小时只需要1美金，而在每秒输出50个 token的情况下， 只需要0.3美金。也就是说，1万美金可以让4个Agent连续工作一年。M2.5提供了经济上几乎无限制地构 建和运营Agent的可能性。对于M2系列的模型来说，唯一的问题变成了模型能力的进步速度。 （文章来源：第一财经） MiniMax微信公众号消息，MiniMax介绍，设计M2系列的初衷是希望能没有成本约束的情况下运行复杂 Agents，随着能力的持续提升，M2.5已经接近实现了这个目标。M2.5提供两个效果一样，但是速度和 价格不一样的版本：速度100TPS左右的快速版本，处理每百万的token输入只需要0.3美金，处理每百万 的token输出只需要2.4美金。而50TPS 的版本的输出价格还更低一倍。按照输出价格参考，50 TPS的版 本价格是Opus、Gemini 3 Pro以及GPT5这些模型的1/10-1/20。 ...

深圳领先边端智能开放研究院揭牌活动上展出的边端智能科技展。 主办方供图 近日，深圳领先边端智能开放研究院(以下简称"研究院")揭牌仪式在深圳市深业上城"城市云客厅"举 行。深圳市政府、有关部门负责人，AI芯片、算法模型、无线通信、终端产品等领域企业、高校和科 研机构代表约40位嘉宾齐聚，共同见证研究院揭牌。 2026年预计将成为"智能体元年" 活动现场，深圳市科技创新局介绍了边端智能战略布局，市政数局和深圳国际科技信息中心分别视频发 布"深小i"政务服务智能体、"AI赋能教科研"智能体最新成果，研究院代表介绍了该院"十五五"规划有关 情况。 采写：南都记者 程洋 深圳市科技创新局局长张林表示，建设深圳领先边端智能开放研究院，是抢抓智能体产业发展机遇、推 进边端智能产业高质量发展的关键举措。边端智能可以有效解决传统云AI在实时性、数据隐私、带宽 成本上的痛点，面向智能终端、智慧家居、可穿戴设备、智能机器人等领域，为智能体尤其是边端智能 体嫁接物理载体与技术底座，对构建自主可控、安全高效的新型智能生态意义重大。 张林表示，2026年预计将成为"智能体元年"，深圳作为全球智能终端第一城，将"边端智能体"纳入城市 发 ...

VAST是2023年3月才成立的年轻企业，自2024年初起持续迭代Tripo大模型，先后推出Tripo1.0至 Tripo3.0等一系列3D大模型，发布了TripoSR、TripoSG、TripoSF等开源3D基础模型，并配套开发了系 列DCC（数字内容创作）软件生态插件。 据央视报道，2月9日，习近平总书记来到位于北京亦庄的国家信创园，开启今年首次国内考察，并同科 研人员和科技企业负责人代表交流。 作为现场科技企业负责人代表之一，VAST创始 兼 席执 官宋亚宸在接受新京报贝壳财经专访时表示， 自己感受到了国家对未来产业和前沿创新的高度重视，也更加坚定了"接过历史接力棒"的决心。 "未来产业是抢占科技制高点、把握发展主动权的关键。"宋亚宸认为，北京的政策红利与"耐心资本"支 持，强大的原始创新能力和密集的人才储备，丰富的概念验证与应用落地场景三大优势深度融合，可以 成为北京加快建设国际科创中心的坚实路径。 他眼中，科技自立自强意味着核心技术必须自主可控，以及创新成果必须赋能产业。在核心技术上 ， VAST押注3D基础大模型，希望通过AI技术重构3D内容创作流程，构建全民化3D内容平台。 "3D抖音"是宋亚宸 ...

Core Insights - Anthropic is now valued at $380 billion, positioning itself alongside OpenAI and SpaceX as one of the most valuable startups to watch for potential public offerings this year [1][4] Funding and Valuation - Anthropic raised $30 billion in its latest funding round, led by Singapore's GIC and U.S.-based Coatue, contributing to its increased valuation [2] - The funding includes part of the $15 billion investment from Nvidia and Microsoft, which is tied to a commitment for Anthropic to purchase $30 billion in computing capacity from Microsoft [3] Business Strategy and Revenue - Anthropic plans to utilize the influx of investments to develop "enterprise-grade products" and AI models [4] - The company is projected to achieve sales of $14 billion over the next year, marking a significant increase from its initial revenue generation less than three years ago [5] Company Background and Focus - Founded in 2021 by former OpenAI employees, Anthropic emphasizes safety in developing artificial general intelligence [6] - The company recently announced a $20 million bipartisan initiative aimed at influencing AI regulation in the U.S. [6] Competitive Landscape - Anthropic is currently the third most valuable private firm, following OpenAI, which is valued at $500 billion, and SpaceX [4]

2月11日晚间，智谱正式确认，此前在全球模型服务平台OpenRouter登顶热度榜首的神秘模型Pony Alpha，为智谱新模型GLM-5。2月12日，智谱在港交所挂牌的股票价格大涨逾28%，收报402港元，本 周以来股价接近翻倍。 农历新年临近，AI科技的"春节档"几乎是连轴转：上周，千问App"一句话点奶茶"吹响了AI从"聊 天"到"办事"的号角；近日，字节跳动的新一代AI视频生成模型Seedance2.0上线，迅速在国内外内容创 作圈刷屏，实测效果被称为"导演级AI"王炸；几乎同一时间，神秘模型Pony Alpha以"免费+强劲性 能"火爆出圈，并推动智谱股价持续飙升。 多款现象级产品密集爆发，不仅点燃了全球AI圈，更成为投资人士热议的焦点。多位接受中国证券报 记者采访的投资人士认为，AI应用商业化正加速落地，除了短期反应强烈的传媒、影视等板块之外， 算力与平台型企业也将长期受益。 业内人士透露，AI视频生成长期受"抽卡式生成"困扰，单个镜头平均需反复生成多次才能获得满意结 果，大幅推高了制作成本与周期。Seedance2.0的核心突破是可控性提升，通过多模态输入实现精准复 刻运镜与动作，减少重复生成， ...

OpenAI has warned US lawmakers that its Chinese rival DeepSeek is using unfair and increasingly sophisticated methods to extract results from leading US AI models to train the next generation of its breakthrough R1 chatbot https://t.co/siq46jRyM1 ...