Core Viewpoint - The article discusses the transition of artificial intelligence (AI) from the digital world to the physical world, marking a critical turning point in the third wave of AI development, with the introduction of the "Wujie" series of large models by the Zhiyuan Institute [12][13][14]. Group 1: AI Development and Trends - The current AI landscape is at a pivotal moment where large models are facilitating the shift from weak AI to general AI, and from specialized robots (1.0) to general embodied intelligence (2.0) [3][13]. - The "Wujie" series of large models aims to bridge the gap between the digital and physical worlds, representing a significant advancement in AI capabilities [4][14]. - The Emu3.5 model, part of the Wujie series, utilizes a unified autoregressive architecture to transition from Next-Token Prediction to Next-State Prediction, indicating a new phase in multimodal learning [17][22]. Group 2: Emu3.5 Model Features - Emu3.5 distinguishes itself by learning from long videos, which contain rich temporal, spatial, and causal information, essential for understanding the physical world [18][20]. - The training dataset for Emu3.5 has significantly expanded, increasing from 15 years to 790 years of video data, and the model parameters have grown from 8 billion to 34 billion [23]. - Emu3.5's autoregressive architecture allows for rapid image generation, achieving speeds comparable to top models through proprietary DiDA technology [23]. Group 3: Multimodal Learning and Applications - Emu3.5 is expected to lead AI into a new stage of multimodal world learning, with substantial scaling potential due to the underutilization of vast multimodal data [24]. - The model demonstrates strong multimodal reasoning and visual understanding capabilities, as evidenced by its performance in image generation and editing tasks [25][27]. - Emu3.5 excels in tasks involving temporal and spatial state predictions, showcasing its superior understanding of the physical world [29][31]. Group 4: Embodied Intelligence and Technological Advancements - The Zhiyuan Institute is addressing the challenges of embodied intelligence, which currently suffers from usability and generality issues [34]. - The institute has developed a comprehensive technology stack centered around the Robo Brain, enabling cross-robot data collection and standardization [35]. - Recent advancements include the RoboBrain2.0, which can decompose complex human instructions for execution by various robots, enhancing the practical applications of embodied intelligence [36]. Group 5: Open Source Contributions - The Zhiyuan Institute has committed to open-source practices, releasing over 200 models and 100 datasets, with global download figures exceeding 690 million and 4 million, respectively [38]. - The institute collaborates with over 30 leading robotics companies to promote the development of embodied intelligence world models [38].

Core Viewpoint - The discussion emphasizes the potential risks and ethical considerations surrounding AI development, particularly in light of recent advancements like ChatGPT, which have raised concerns about AI becoming a competitive entity to humans and the implications for society [6][7][9]. Group 1: AI Risks and Responsibilities - Bengio acknowledges the responsibility of researchers in the AI field for the potential risks associated with their work, highlighting a personal emotional shift towards recognizing these dangers after the emergence of ChatGPT [10][12][13]. - The probability of catastrophic outcomes from AI, even at a low percentage, is deemed unacceptable, urging for increased societal attention and investment in AI safety [17][22]. - The divergence in expert opinions regarding AI risks indicates a lack of sufficient information to predict future outcomes, suggesting that pessimistic views may hold validity [20][21]. Group 2: AI's Impact on Employment - AI is expected to replace many cognitive jobs in the near future, while physical jobs, such as plumbing, may remain unaffected temporarily due to current limitations in robotics technology [50][48]. - The integration of AI into workplaces is driven by companies' motivations to enhance efficiency and profitability, despite the potential for significant job displacement [50][53]. Group 3: Ethical Considerations and Future Directions - The conversation stresses the importance of ethical AI development, advocating for a shift from profit-driven motives to a focus on societal well-being and safety [44][80]. - There is a call for global cooperation to manage the risks associated with AI, particularly as it becomes more integrated with robotics and other technologies that could pose physical threats [56][62]. - The need for public awareness and understanding of AI risks is emphasized, suggesting that individuals should educate themselves and engage in discussions about AI's implications [83][89].

允中 发自 凹非寺 量子位 | 公众号 QbitAI 近日， 深势科技完成总额超8亿人民币的C轮融资 ，本轮融资由达晨财智、京国瑞基金、北京市人工智能产业投资基金、北京市医药健康产业 投资基金、联想创投、元禾璞华等机构共同出资。 本轮融资资金将主要用于持续吸引和培养行业内顶尖人才，进一步进化迭代深势科技的"科学发现智能引擎"，持续夯实从原始技术创新、到智 能科研工具产品及行业解决方案的全栈能力，加速围绕科学发现的智能产品与服务在 基础科研、生命科学与物质科学 等领域的市场拓展与规 模化应用。 此次融资的完成，标志着深势科技在构建新一代科学发现智能引擎的征程上，迈出了坚实的一步。 AI for Science成为全球共识，科学发现范式正在重构 我们正站在一个历史性的时点，AI for Science已成为全球性的共识，其目标在于从根本上变革人类探索未知、发现全新科学知识并将其系统 沉淀为可复用科学资产的模式。 2025年8月，国务院发布关于深入实施"人工智能+"行动的意见，意见将 "人工智能+科学研究" 放在首位，其中特别强调了加速科学发现进 程，驱动技术研发模式创新和效能提升。 此外，欧洲的"地平线"计划中重 ...

Core Viewpoint - The article emphasizes the ongoing AI boom and invites individuals to join the company "Quantum Bit," which focuses on tracking AI advancements and has established itself as a leading content platform in the industry [1]. Group 1: Job Opportunities - The company is hiring for three main directions: AI Industry, AI Finance, and AI Product, with positions available for both experienced professionals and fresh graduates [2][4]. - Positions are full-time and based in Beijing, with various levels of roles open for application [2][4]. Group 2: Job Responsibilities - **AI Industry Direction**: Focuses on innovations in infrastructure, including chips, AI infrastructure, and cloud computing [6]. - **AI Finance Direction**: Involves tracking venture capital and financial reports in the AI sector, monitoring capital movements within the industry [6]. - **AI Product Direction**: Concentrates on the application and hardware advancements in AI [6]. Group 3: Benefits and Growth Opportunities - Employees will have the chance to engage with the latest AI technologies, enhance their work efficiency through new AI tools, and build personal influence by creating original content [6]. - The company offers competitive salaries, comprehensive benefits including social insurance, meal allowances, project performance bonuses, and a supportive team environment [6]. Group 4: Company Achievements - As of 2025, Quantum Bit has over 2.4 million subscribers on WeChat and more than 7 million users across platforms, with a daily reading volume exceeding 2 million [12]. - The company is recognized as the top new media outlet in the AI and frontier technology sector according to third-party data platforms [12].

Core Insights - The article discusses Tencent's algorithm competition, highlighting its significance in attracting talent and providing practical experience in cutting-edge AI technologies [1][28][43] Group 1: Competition Overview - The competition offered substantial rewards, including a total prize pool of 3.8 million yuan, with the champion receiving 2 million yuan and all participants gaining access to valuable resources like computing power [32][34] - The competition attracted over 8,400 students and 2,800 teams from nearly 30 countries, showcasing its global reach and influence [34] Group 2: Technical Focus - The competition's theme, "full-modal generative recommendation," addresses advanced challenges in advertising and recommendation systems, emphasizing the integration of various data types such as text, images, and videos [5][11] - Participants faced real-world challenges, including data noise, alignment issues, and the need for efficient modeling of user behavior over long sequences [13][41] Group 3: Talent Acquisition Strategy - Tencent's approach to the competition serves as a recruitment strategy, allowing the company to identify and engage with top talent in a practical setting rather than traditional recruitment methods [39][42] - The competition's structure inherently filters candidates, ensuring that only those capable of handling complex data and modeling challenges progress to the final stages [40][41] Group 4: Industry Context - The competition reflects Tencent's established AI technology framework, which has been validated through real business applications, indicating the company's commitment to innovation and talent development [29][30] - The article notes the competitive landscape for talent in the AI sector, with companies like Tencent offering attractive employment packages and support programs to attract young professionals [44][46]

Core Viewpoint - The article discusses a heated debate between AI experts Yann LeCun and Demis Hassabis regarding the nature of intelligence, particularly focusing on the concept of "general intelligence" and its implications for artificial intelligence development [3][8][30]. Group 1: Debate Overview - Yann LeCun argues that the idea of "general intelligence" is nonsensical, asserting that human intelligence is highly specialized rather than universal [9][13]. - Demis Hassabis counters LeCun's claims, stating that human brains exhibit significant generality and complexity, and that general intelligence is a valid concept [17][22]. - The debate has attracted considerable attention, with notable figures like Elon Musk publicly supporting Hassabis [5][7]. Group 2: Key Arguments - LeCun emphasizes that human intelligence is shaped by evolutionary pressures to adapt to specific environments, leading to specialized skills rather than general capabilities [14][36]. - Hassabis argues that the brain functions similarly to a Turing machine, capable of learning any computable content given sufficient resources, thus supporting the existence of general intelligence [18][24]. - The discussion highlights a fundamental disagreement over terminology, with LeCun focusing on the specialized nature of human cognition while Hassabis advocates for the potential of general intelligence [32][41]. Group 3: Future Directions in AI - Both experts agree on the importance of "world models" in advancing artificial general intelligence (AGI), though they have different interpretations of what this entails [42][50]. - LeCun's upcoming venture, Advanced Machine Intelligence Labs, aims to develop world models that prioritize understanding control theory and cognitive science [43][44]. - Hassabis and Google DeepMind are also focusing on world models, emphasizing the need for models that comprehend causal relationships and interactions within the world [46][47].

梦晨 发自 凹非寺 量子位 | 公众号 QbitAI "从小就要赢在起跑线" 这套逻辑，被顶刊Science最新论文狠狠打了脸。 这项研究综合分析了超过34000名国际顶尖人才的成长轨迹，涵盖诺贝尔奖得主、典作曲家、奥运冠军以及世界顶级棋手。 结论颠覆人们观念： 作者团队来自德国凯泽斯劳滕工业大学 （RPTU Kaiserslautern） 体育科学系、密歇根州立大学心理学系、普渡大学心理科学系。 他们综合分析了多项研究数据，涵盖科学、艺术、体育多个领域。 少年天才往往止步于顶尖水平之下，和最终登顶的成年人近90%不是同一批人。 而最终达到世界级水平的人才，在早年阶段表现反而低于只达到国家级水平的同龄人。 "天才少年"长大后去哪了 长久以来，学界对人才培养的研究主要聚焦于年轻人。传统观点普遍认，早期表现越好、专项练习越多，后期成就越高。 全球各地的精英学校、音乐学院和青训学院也据此设计了选拔机制：挑出表现最好的年轻人，然后用高强度的专项训练进一步"加速"他们的成 长。 但这套逻辑在真正的世界顶尖群体中是否成立，此前从未被系统验证过。 通过大规模数据追踪，研究团队给出了一个令人意外的答案：无论是体育、国际象棋还 ...

Core Insights - The article emphasizes that 2025 marks a significant year for AI advancements, particularly in reasoning, collaboration, and scientific discovery, led by Google [1][3][9] Group 1: AI Development and Integration - Google has made substantial progress in reasoning, multi-modal understanding, model efficiency, and generative capabilities, significantly enhancing model performance [15][4] - The Gemini series, particularly Gemini 3 Pro, has set new standards in multi-modal reasoning and achieved top scores in various benchmark tests, including a 23.4% record in MathArena Apex [18][19] - AI has been deeply integrated into Google's core products, transforming from a tool to a practical asset for users [5][10][23] Group 2: Generative Media and Creative Tools - 2025 is highlighted as a transformative year for generative media, with AI providing unprecedented capabilities for video, image, audio, and virtual world generation [24][25] - Google has collaborated with creative professionals to develop tools like Flow and Music AI Sandbox, enhancing creative workflows [25][21] Group 3: Scientific and Mathematical Advancements - AI has significantly contributed to advancements in life sciences, health, natural sciences, and mathematics, empowering researchers with new tools and resources [27][28] - The AI system AlphaFold, which addresses protein folding, has been widely adopted by researchers globally, marking a milestone in scientific research [28] Group 4: Quantum Computing and Physical World Research - Google has made notable advancements in quantum computing and energy-efficient technologies, including the launch of a new TPU designed for the reasoning era [33][32] - The company has also made strides in robotics and visual understanding, integrating AI agents into both physical and virtual environments [33] Group 5: Addressing Global Challenges - Google's AI-driven scientific progress is being applied to tackle critical global challenges, including climate resilience, public health, and education [36][38] - The company has developed advanced forecasting models that enhance decision-making in various sectors, including weather prediction [36] Group 6: Responsibility and Safety - Google emphasizes the importance of combining research breakthroughs with responsibility and safety, continuously improving tools and frameworks to mitigate risks [42][43] - The Gemini 3 model is noted as the safest model to date, undergoing comprehensive safety assessments [44] Group 7: Collaboration and Open Ecosystem - Google advocates for cross-sector collaboration to responsibly advance AI, establishing partnerships with leading AI labs and educational institutions [46][45] - The company aims to continue promoting cutting-edge technology safely and responsibly for the benefit of humanity [47]

克雷西 发自 凹非寺 量子位 | 公众号 QbitAI MiniMax最新旗舰级Coding & Agent模型 M2.1 ，刚刚对外发布了。 一边是港交所聆讯通过新进展，另一边新模型还在嗖嗖嗖上新——而且还SOTA了。 这一次，它直接甩出了一份硬核成绩单，在衡量多语言软件工程能力的Multi-SWE-bench榜单中，以仅10B的激活参数拿下了49.4%的成绩， 超越了Claude Sonnet 4.5等国际顶尖竞品，拿下全球SOTA。 它试图解决的，就是此前模型身上严重的"学科偏科"问题。 所谓偏科，指的是过去的模型，写写Python脚本或Web前端页面表现还可以，可一旦涉及到后端架构，亦或底层逻辑，表现往往会出现断崖 式下跌。 M2.1的核心进化，就在于它终于突破了这个难题，掌握了后端的开发规范。 M2.1的发布，也证明了MiniMax在推进上市流程的同时，仍保持着高频的研发节奏。 更懂底层，10B激活参数拿下SOTA M2.1将对工程上下文的理解，转化为了对开发工具链的深度适配。它不仅能生成代码，更能熟练配合Cursor、Claude Code等主流编程工 具，在存量代码库中执行精准的修复（Fix）或 ...

鹭羽 发自 凹非寺 量子位 | 公众号 QbitAI 我真栓Q了！围观了场 狼人杀 ，看得我汗流浃背…… 半小时全程高能，根本停不下来： 天崩开局倒钩狼悍跳预言家、冲锋狼死于话多、神职上大分每晚都是平安夜。 结果你跟我说，这些玩家都是 AI ？？？ 果然会玩还得看 淘宝 ～最近他们整活的这个AI狼人杀大乱斗 WhoisSpy.ai ，大模型在里面简直咔咔乱杀。 D老师、Qwen、Kimi、GLM一个个都化身心机boy推拉博弈，be like： …… 不过u1s1，虽然这些Agent看似性格迥异，实则一个个都是狼人杀高玩来着。 而且门槛也不高，自己就能手搓一个出来。 是不是有点手痒了？ （咳咳） 不卖关子了，这就是我最近刷到的一个AI狼人杀比赛，还是淘宝办的——首届 「高校生VS开发者对抗赛」 。 展开来说，就是淘宝发了个召集令，广邀高校学生和AI开发者，带着自家Agent来真刀实枪碰一场，看看谁的Agent思维更缜密、更会盘逻 辑。 六边形战士 Kimi ：武力值MAX，第六感Next Level。 老实人 DeepSeek ：虽然我只是一介平民，虽然我只会划水，但我相信跟对人走对路，奥利给！ 喜剧人 Qwe ...