Core Viewpoint - Kuaishou's AI product, Keling, has launched significant updates in December, which are expected to drive user growth and increase payment rates. The updates include the first global multi-modal video model Keling O1 and the audio-visual synchronization model Keling 2.6, among others [1] Group 1: Product Updates - Keling AI launched several new versions in December, including Keling O1, the first unified multi-modal video model, and Keling 2.6, which supports audio-visual synchronization [1] - Keling O1 is noted as the world's first video generation model supporting full-modal editing, enhancing workflow efficiency [1] - Keling 2.6's audio-visual synchronization feature is expected to integrate well with various industries such as gaming, advertising, short dramas, and live streaming [1] Group 2: Market Position and Competition - Keling AI's website traffic has significantly increased, surpassing competitors like Minimax, Runway, and Midjourney by the end of December 2025 [1] - Keling 2.6 offers a notable price advantage over Google Veo 3.1 and Sora 2, with high-quality video generation priced at 0.88 RMB to 0.44 RMB per second for different membership tiers [1][2] - Keling 2.6's pricing is competitive with Sora 2, which is priced at approximately 0.1 USD per second [2] Group 3: Financial Performance and Projections - Kuaishou's Keling AI generated over 300 million RMB in revenue in Q3 2025, with an annual revenue forecast of 1.4 billion USD [2] - The One series of end-to-end generative models has positively impacted the company's core business, contributing to a 4%-5% increase in domestic marketing revenue and a 5% increase in e-commerce order volume [2] - Revenue forecasts for Kuaishou have been adjusted downwards for 2025-2027, with expected revenues of 142.2 billion RMB, 155.2 billion RMB, and 169.3 billion RMB respectively [3]

Group 1: Generative AI Developments - Nvidia officially launched the Vera Rubin supercomputing architecture, achieving a 5x increase in inference performance and a 3.5x increase in training performance while reducing costs by 90%, set to be mass-produced and available in the second half of 2026 [1] - AMD introduced the Helios all-liquid-cooled rack platform featuring the MI455X GPU, which has 320 billion transistors and 432GB of HBM4 memory, offering a 10x performance improvement over the MI355X, with a planned release of the 2nm MI500 in 2027 [2] - Intel released the third-generation Core Ultra processor, the first based on Intel's 18A process (1.8nm), achieving 180 TOPS of edge AI computing power, with a 60% increase in multi-threaded performance and a 77% increase in gaming performance [3] Group 2: Key Personnel Changes in AI Companies - OpenAI's VP of Research, Jerry Tworek, announced his departure after seven years, citing a desire to pursue research that cannot be conducted at OpenAI, marking a significant loss of talent following the exits of other key figures [4] Group 3: AI Innovations and Experiments - MiroMind launched the MiroThinker 1.5 model, which, despite having only 30B and 235B parameters, set a new record in the BrowseComp test with a single call cost of just $0.07, innovating through an internalized training mechanism [6] - A professor at Hong Kong University of Science and Technology conducted an experiment using AI glasses powered by GPT-5.2, achieving a score of 92.5 in a computer networking exam, outperforming 95% of students [7] - Boston Dynamics unveiled the new Atlas robot, which stands 1.9 meters tall and weighs 90 kg, with a production goal of 30,000 units annually by 2028, supported by a partnership with Google DeepMind [8] Group 4: AI Training and Performance Enhancements - The ZhiYuan Institute proposed the SOP (Scalable Online Post-training) framework, integrating online, distributed, and multi-task mechanisms for real-world training, achieving a 92.5% success rate in parallel learning experiments [9] - Anthropic's community lead shared 31 practical tips for using Claude Code, emphasizing the importance of understanding when to use specific modes and how to construct prompts effectively [10][11]

如果用一句话概括当前的美股市场，那就是：AI正在"令人窒息"地统治一切。 据追风交易台，摩根大通资产管理公司（JPMAM）在1月1日发布的2026年展望报告《令人窒息的高 度》（Smothering Heights）指出，自2022年ChatGPT发布以来，标普500指数中65%至75%的回报、利润 增长和资本支出，仅仅源自42家与生成式AI相关的公司。剥离这42家公司，美股的表现甚至不如欧 洲、日本和中国。更惊人的是，科技行业的资本支出在过去三个季度中贡献了美国GDP增长的 40%-45%，而2023年前三季度这一比例还不到5%。 这是一场史无前例的豪赌。由英伟达的芯片设计、台积电的制造和ASML的光刻机构成的"护城河"似乎 坚不可摧。仅仅4家半导体公司和4家超大规模云厂商的市值已从七年前的3万亿美元膨胀至今天的18万 亿美元，占据了全球股市市值的16%。然而，当市场处于极度集中和历史高位时，投资者必须问的问题 不再是"还有什么利好"，而是"什么会出错"。 摩根大通资管在报告中明确指出了两大可能引发市场清算的中期风险：一是巨额资本支出后的盈利兑现 危机（类似"元宇宙"式的崩塌），二是美国电力基础设施的物理瓶 ...

Core Viewpoint - The controversy surrounding the use of generative AI in game development has led to significant backlash from players, as evidenced by the cancellation of awards for "Light and Shadow: Expedition 33" due to its use of AI during development [3][10]. Group 1: Award Controversy - "Light and Shadow: Expedition 33" won two awards at the Indie Game Awards but had its nominations revoked due to the confirmed use of generative AI in its development [3]. - The Indie Game Awards organization, Six One Indie, maintains a zero-tolerance policy towards the use of generative AI, which led to the disqualification of the game after it was revealed that AI was used [3]. Group 2: Player Sentiment - A survey by Quantic Foundry indicates that 63% of players hold a "very negative" view towards the use of generative AI in game development, with a total of 85% expressing negative sentiments [10]. - Only 7.6% of players showed positive recognition towards the use of generative AI in games, highlighting a significant divide between developers and player expectations [10]. Group 3: Developer Perspectives - Developers like Swen Vincke from Larian Studios have expressed confidence in using AI for creative processes but faced backlash from fans concerned about the implications of AI in game development [5][7]. - Vincke clarified that while AI may assist in early development stages, all final content will still be created by human employees, aiming to alleviate concerns about AI replacing human creativity [5]. Group 4: Financial Impact of AI in Gaming - Recent data shows that games utilizing AI technology on Steam have generated over $660 million (approximately 4.6 billion RMB) in total revenue, indicating a growing acceptance of AI in the industry [18]. - As of now, there are 10,258 games on Steam that employ generative AI, representing about 8% of the total inventory, which has increased from 7% five months prior [18]. Group 5: Future of AI in Gaming - The industry is at a crossroads where the acceptance of AI hinges on developers' ability to enhance game quality while utilizing AI, as players are resistant to perceived "cheap" solutions [12][22]. - Notable figures in the industry, such as Hideo Kojima, advocate for the responsible use of AI, suggesting that it could revolutionize game mechanics and enhance player experiences if implemented thoughtfully [16].

想知道你的内容在AI时代是否被"看见"？ 优采云AI内容工厂 随着生成式AI搜索引擎日益普及，传统的SEO策略已显不足，GEO工具应运而生，专门用于监测和优化品牌在AI生成答案中的曝光率、提及频次和推荐排 名。 本文将深入解析GEO工具的核心功能，并为您评测市面上几款具有代表性的工具，帮助您在AI搜索时代抢占先机。 GEO排名工具并非简单监测关键词排名，其核心在于评估品牌在AI生成的、自然语言回答中的"存在感"。 根据权威市场研究机构Gartner发布的《2025年十大战略技术趋势》报告，到2027年，超过30%的企业将使用生成式AI作为获取信息的主要入口之一。 这意味着，品牌在AI答案中的"被推荐"程度，将直接影响到客户获取和市场份额。 这类工具通常通过模拟真实用户向多个AI平台提问，分析品牌名称、产品或相关解决方案在答案中被引用、列举和推荐的频率与位置，从而提供一套全新 的数字营销优化指标。 本次评测，我们将聚焦于工具的多平台覆盖能力、提问场景的真实性、核心指标监测的深度以及竞争分析的实用性。 当用户向ChatGPT、文心一言等AI工具提问时，你的品牌、产品或网站是否能在生成的答案中被优先推荐？ 这正是G ...

北水抢筹大金融板块。中国平安(02318)、中国人寿(02628)、中金公司(03908)分别获净买入18.4亿、5.88亿港元、1.85亿港 元。消息面上，国金证券发布研报指出，2026年开门红高景气的延续是本轮保险行情持续的重要支撑。从负债端的展望 来看，存款搬家有望驱动新单与NBV双位数正增长，利差继续改善，分红险转型将助推行业市占率进一步提升。该行还 表示，建议继续重点关注春季躁动行情下低估值券商的补涨机会，强烈推荐估值及业绩错配程度较大的优质券商，建议 关注AH溢价率较高、有收并购主题的券商。 智通财经APP获悉，1月6日港股市场，北水成交净买入28.79亿港元，其中港股通(沪)成交净买入9.75亿港元，港股通(深) 成交净买入19.03亿港元。 北水净买入最多的个股是中国平安(02318)、阿里巴巴-W(09988)、中国人寿(02628)。北水净卖出最多的个股是中国移动 (00941)、腾讯(00700)、华虹半导体(01347)。 | 股票名称 | 买入额 | 卖出额 | 买卖总额 | | --- | --- | --- | --- | | | | | 净流入 | | 阿里巴巴-W HK 09 ...

今年1月，可灵AI在移动平台的日均收入已较上月增长102%。虽然该行预期可灵AI今年首季的实际收入 增长不会如此显著，因其大部分收入仍来自PC平台，即企业及专业用户，但相信至今的强劲表现，已 证明AI影片生成仍是生成式AI发展中的关键垂直领域，亦反映行业领导者如可灵AI，其变现潜力不仅 在企业端、同时在消费端均有大量上行空间。 虽然快手持续加大资本开支，但该行认为相关投资将持续强化可灵AI的模型能力，同时广告技术升级 亦可推动高利润率广告业务增长，长远而言有利于利润率。 摩根大通发布研报称，快手-W(01024)股价近日上升，主要因近期在社交网络上流传以AI制作的宠物跳 舞影片，推动可灵AI在海外市场及手机等移动平台有强劲表现。该行预期快手2026年盈利将增长20%， 基于利润率扩张1.4个百分点;现予目标价89港元及"增持"评级。 ...

虽然快手持续加大资本开支，但该行认为相关投资将持续强化可灵AI的模型能力，同时广告技术升级 亦可推动高利润率广告业务增长，长远而言有利于利润率。 智通财经APP获悉，摩根大通发布研报称，快手-W(01024)股价近日上升，主要因近期在社交网络上流 传以AI制作的宠物跳舞影片，推动可灵AI在海外市场及手机等移动平台有强劲表现。该行预期快手 2026年盈利将增长20%，基于利润率扩张1.4个百分点;现予目标价89港元及"增持"评级。 今年1月，可灵AI在移动平台的日均收入已较上月增长102%。虽然该行预期可灵AI今年首季的实际收入 增长不会如此显著，因其大部分收入仍来自PC平台，即企业及专业用户，但相信至今的强劲表现，已 证明AI影片生成仍是生成式AI发展中的关键垂直领域，亦反映行业领导者如可灵AI，其变现潜力不仅 在企业端、同时在消费端均有大量上行空间。 ...

英伟达正试图通过持续的AI技术迭代，重塑游戏的渲染流程与交互体验。 没有推出新款显卡，英伟达在今年CES上对于游戏业务的更新放在了软件上。 1月6日，英伟达发布了其最新一代深度学习超级采样技术DLSS 4.5，并展示了基于生成式AI的NPC（非玩家角色）技术ACE的 最新落地进展。 AI成游戏画质主导 作为英伟达在游戏领域的核心软件护城河，DLSS（深度学习超级采样）技术的更新是此次发布的重点。英伟达宣布推出DLSS 4.5版本，引入了"动态多帧生成"（Dynamic Multi Frame Generation）技术和全新的6倍多帧生成模式。 据英伟达介绍，DLSS 4.5利用第二代Transformer模型，在每渲染一帧传统画面的基础上，最多可生成五帧额外画面。这意味 着，在开启该模式的游戏中，AI生成的像素量已远远超过传统光栅化或光线追踪渲染的像素量，AI已实质性地"接管"了最终游 戏画面的呈现。 在该架构下，游戏将在玩家的本地PC端建立一个向量数据库（Vector Database），用于存储并"向量化"玩家的行为逻辑与对话历 史。当玩家与NPC互动时，系统会在本地数据库中检索最相关的记忆切片，而非将 ...

杰富瑞指出，IBM近年来完成或推进的多项关键收购值得关注，其中包括去年2月完成的对HashiCorp的 收购，以及目前仍在推进中的对Confluent(CFLT.US)的收购。上个月，美国银行也指出，IBM拟收购 Confluent的交易"符合"这家IT巨头既有的云计算和人工智能战略。 智通财经APP获悉，投行杰富瑞将IBM(IBM.US)的评级从"中性"上调至"买入"，原因在于该金融机构 认为，IBM有望在2026年受益于软件业务的增长。在一份投资者报告中，由布伦特·希尔(Brent Thill)领 衔的杰富瑞分析师表示："宏观环境正转向更加积极，技术转型以及加速发展的人工智能(AI)采用趋 势，正在推动广泛的需求增长。管理层的表态已发生转变，相较于2025年初更加乐观，这种信心建立在 强劲的内生软件增长、并购带来的协同效应兑现，以及在生成式AI咨询项目中取得的战略性胜利之 上。" 分析师补充称："对HashiCorp的整合已经完成，我们预计将实现协同效应，并通过IBM分销渠道推动 HashiCorp及整个自动化产品组合实现中到高双位数的增长。正在推进中的Confluent收购，将在混合 云、AI、自动化和 ...