7月28日晚，清华系大模型明星企业智谱发布新一代旗舰模型GLM-4.5，据悉，这一基础模型专为智能 体应用打造，已在Hugging Face与ModelScope平台同步开源。 据悉，GLM-4.5兼容Claude Code、Cline、Roo Code等主流代码智能体，海内外用户可以在智谱开放平 台体验。 来源：北京日报客户端 记者：孙奇茹 如遇作品内容、版权等问题，请在相关文章刊发之日起30日内与本网联系。版权侵权联系电话：010-85202353 技术人员介绍，GLM-4.5 参数量为 DeepSeek-R1的二分之一、Kimi-K2 的三分之一，但在多项标准基准 测试中表现得更好，得益于模型更高的参数效率。在性能优化之外，GLM-4.5系列也在成本和效率上实 现突破：API 调用价格低至输入价格为0.8元/百万tokens，输出价格2元/百万tokens，大幅低于目前主流 的模型定价。 "衡量AGI（通用人工智能）的第一性原理，是在不损失原有能力的前提下融合更多通用智能能力， GLM-4.5 是我们对此理念的首次完整呈现，并有幸取得技术突破。"智谱相关负责人介绍，GLM-4.5 首 次在单个模型中实现 ...

Group 1: Starship Development - The Starship project is considered one of the most ambitious rocket endeavors, with challenges anticipated at every stage, but no unexpected difficulties have arisen [1][2] - Starship's thrust is 2.5 times that of the Saturn V rocket, with future versions expected to increase this to 3 times, aiming for complete and rapid reusability [1][2] - The most significant current challenge is developing a fully reusable thermal protection shield, a technology that has never been achieved before [2][3] Group 2: Mars City Vision - The governance of Mars cities will be determined by the residents, with initial living conditions requiring protective gear until terraforming is achieved [6] - The rationale for establishing a self-sustaining city on Mars includes both defensive and inspirational values, ensuring human survival in case of Earth disasters and providing a grand goal for humanity [6] Group 3: Robotaxi and Tesla's Electric Vehicle Strategy - The Cybercab will not replace Model 3 and Model Y, as it is designed for 1-2 passengers, while the latter models cater to larger groups [8] - Tesla plans to maintain a portion of the Robotaxi fleet while allowing private owners to join the shared network, similar to a combination of Uber and Airbnb [8][9] Group 4: Optimus Robot Potential - The third version of the Optimus robot is designed for mass production, with a potential market size estimated between $20 billion and $500 billion [10] - Initial applications of Optimus will focus on high-value scenarios, such as providing assistance to individuals with severe disabilities [12] Group 5: AI and Grok's Mission - The development of AI is progressing rapidly, with concerns about ensuring that AI aligns with human interests and remains beneficial [14][15] - The integration of Grok with Neuralink aims to enhance communication efficiency by allowing direct transmission of thoughts without language conversion [18] Group 6: Neuralink's Long-term Vision - Neuralink's primary goal is to assist individuals with severe disabilities, with plans to expand user trials and introduce new technologies for vision restoration [17][19] - The potential for Neuralink to enhance human intelligence and communication speed is seen as a significant future development [17][19] Group 7: X Platform Transformation - The X platform is evolving to support longer videos and texts, aiming to ensure freedom of speech while correcting misinformation through community notes [21][22] - The platform's algorithm is being optimized to improve user experience and content relevance [21][22] Group 8: Human Role in an AI-Driven Future - The integration of AI and robotics may lead to a significant expansion of the economy, but it raises philosophical questions about the meaning of human existence [23][24] - The goal is to enhance human capabilities through technology while ensuring that humans retain their significance in a future dominated by AI [23][24]

Group 1: AI Innovations in Production and Daily Life - AI is leading a new technological revolution, enhancing productivity and creativity in various fields, such as video generation and 3D modeling [2][3] - The "Keling AI" platform allows users to create videos from text inputs, significantly improving creative efficiency and freeing up creators' productivity [3] - Smart glasses showcased at the conference enable users to interact and place orders in different languages, demonstrating practical applications of AI in everyday scenarios [5] Group 2: AI's Impact on Scientific Research - AI is reshaping the foundational logic of scientific research, accelerating discoveries and providing new opportunities to solve significant scientific challenges [6] - Notable advancements include AI's ability to predict protein folding structures and assist in various scientific fields such as quantum computing and cancer detection [6][7] - The "Panshi Scientific Foundation Model," developed by the Chinese Academy of Sciences, aims to enhance scientific tasks through deep understanding of scientific data and literature [6][7] Group 3: Global Cooperation and Governance in AI - The development of AI is recognized as a global issue that requires collaborative governance and innovation to maximize its benefits across industries [8] - The "Global Governance Action Plan for Artificial Intelligence" was introduced, emphasizing the need for cooperation in AI innovation, safety governance, and equitable development [8] - China is actively working to bridge the intelligence gap and promote shared benefits through AI applications in education, agriculture, and healthcare across various regions [9]

Core Viewpoint - The article discusses the shift in AI research from "can it be done" to "is it effective," emphasizing the need for new evaluation methods that assess the long-term adaptability and learning capabilities of models, particularly in the context of achieving general artificial intelligence [1][4]. Group 1: New Evaluation Paradigm - A new evaluation paradigm called EvaLearn has been proposed to assess the learning ability and efficiency of large language models (LLMs), providing a fresh perspective on understanding their human-like learning potential [5][6]. - EvaLearn focuses on "sequential problem-solving," redefining the evaluation logic for large language models, and has gained significant attention since its open-source release [6][8]. Group 2: Limitations of Traditional Benchmarks - Traditional benchmarks treat problems as isolated samples, failing to evaluate models' learning efficiency and adaptability, which are crucial for understanding their performance [8][9]. - EvaLearn constructs 648 challenging problems organized into 182 sequences, requiring models to solve them in order, thus allowing for a systematic assessment of their learning capabilities [9][11]. Group 3: Key Findings from EvaLearn - The research team found that models exhibit diverse learning abilities across different task types, with most models better leveraging prior experience for mathematical and logical reasoning tasks, while tasks like summarization rely more on pre-trained knowledge [14]. - Models based on chain-of-thought reasoning generally outperform those that are not, demonstrating better stability and the ability to solve multiple related problems consecutively [15]. - Feedback learning, which incorporates evaluations from a verifier, significantly enhances models' learning abilities and efficiency compared to example-based learning [16]. - Learning ability and efficiency metrics provide a comprehensive assessment of models' learning potential, revealing that high static performance does not guarantee superior learning capabilities [17]. Group 4: Evaluation Metrics - EvaLearn employs a comprehensive set of evaluation metrics to characterize models' dynamic learning abilities, including summary accuracy, classification skills, information extraction, logical reasoning, mathematical reasoning, and sequence reasoning [20]. - Overall accuracy, learning speed, first correct position, consecutive correct answers, and post-warm-up accuracy are key indicators used to assess models' performance [21]. Group 5: Learning Efficiency and Methods - The study indicates significant differences in learning efficiency among models and task types, with non-thinking models often showing faster progress in experience accumulation, while thinking models demonstrate more stable gains [44]. - Different problem-solving methods, such as example learning and feedback learning, significantly impact model performance, with feedback learning generally yielding higher accuracy and learning efficiency [46][48]. - The average position of the first correct answer varies across models and tasks, highlighting the models' learning potential and the importance of feedback in enhancing learning outcomes [51][53]. Group 6: Conclusion - EvaLearn represents a novel benchmark framework for sequentially evaluating models' learning abilities and efficiencies across various tasks, revealing significant performance differences among leading models [55][56]. - The findings underscore the importance of understanding models' learning capabilities and efficiencies as a new perspective for evaluating their performance and bridging the gap between current models and human capabilities [57].

Core Insights - Jumpshare Star announced three significant developments: the launch of the new generation foundational model Step3, a strategic partnership with Shanghai State-owned Capital Investment Co., and the establishment of the Model Ecological Innovation Alliance with nearly ten chip manufacturers and computing power platforms [2][3][6]. Group 1: New Model Launch - The new foundational model Step3 is designed to balance intelligence and efficiency, aiming to create the most suitable model for the inference era and will be open-sourced to global enterprises and developers on July 31 [3]. - Step3 boasts a decoding efficiency that can reach up to 300% compared to DeepSeek-R1 on domestic chips, and it is compatible with all chip types [3]. Group 2: Strategic Partnership - The collaboration with Shanghai State-owned Capital Investment Co. marks a significant step in the commercialization of Jumpshare Star, focusing on capital linkage, ecosystem construction, business synergy, and application empowerment [6][9]. - Shanghai State-owned Capital Investment Co. has a registered capital of 10 billion yuan and is involved in strategic equity management and market-oriented investment projects [9]. Group 3: Ecosystem Alliance - The Model Ecological Innovation Alliance aims to enhance model adaptability and computing efficiency through collaborative innovation among foundational technology vendors [11]. - Initial members of the alliance include major companies such as Huawei Ascend, MuXi, and others, with the goal of providing efficient and user-friendly large model solutions [11][13].

Core Viewpoint - Jumpshare Star announced three significant developments: the launch of the new generation foundational model Step 3, a strategic partnership with Shanghai State-owned Capital Investment Co., and the establishment of the Model Ecological Innovation Alliance with nearly ten chip manufacturers and computing power platforms [1][7][14]. Group 1: New Generation Foundational Model - The new foundational model Step 3 is designed to balance intelligence and efficiency, aiming to create the most suitable model for the inference era and contribute a powerful multimodal inference model to the open-source community [1][2]. - Step 3 achieves a decoding efficiency that is up to 300% higher than DeepSeek-R1 on domestic chips, demonstrating significant advancements in system and architecture innovation [2]. - In distributed inference using NVIDIA Hopper architecture chips, Step 3 shows a throughput improvement of over 70% compared to DeepSeek-R1 [4]. Group 2: Strategic Partnership - The partnership with Shanghai State-owned Capital Investment Co. marks a significant step in Jumpshare Star's commercialization efforts, focusing on capital linkage, ecological construction, business collaboration, and application empowerment [7]. - Shanghai State-owned Capital Investment Co. is a large state-owned capital investment platform with a registered capital of 10 billion yuan, involved in strategic equity management and market-oriented investment [8]. Group 3: Commercialization Progress - Jumpshare Star has achieved commercial progress, with over half of domestic smartphone manufacturers collaborating with the company, and partnerships with Geely Auto for smart cockpit solutions [10]. - The company aims to achieve an annual revenue target of 1 billion yuan by 2025, driven by rapid growth in the first half of 2025 [10]. Group 4: Model Ecological Innovation Alliance - The Model Ecological Innovation Alliance, initiated by Jumpshare Star and nearly ten chip and infrastructure manufacturers, aims to enhance model adaptability and computing efficiency through collaborative innovation [14][15]. - Initial members of the alliance include Huawei Ascend, Muqi, and several other technology firms, with the goal of providing efficient and user-friendly large model solutions for enterprises and developers [14][15].

Group 1 - Meta has appointed Shengjia Zhao, a Tsinghua University alumnus, as the Chief Scientist of its Superintelligent Lab (MSL) [1][2] - Mark Zuckerberg expressed excitement about Zhao's leadership and his groundbreaking contributions in various fields, emphasizing the formation of a high-density talent team for long-term development [4][28] - Zhao has a strong academic background, having graduated from Tsinghua University and earned a PhD from Stanford University, focusing on large model architectures and multimodal reasoning [12][21] Group 2 - Zhao was a core member at OpenAI, significantly contributing to the design of GPT-4 and other models, and has been involved in key technical paths such as model reasoning and safety mechanisms [15][17] - His work includes being a primary author of the highly cited "GPT-4 Technical Report," which has over 17,000 citations, making it one of the most referenced documents in contemporary AI [18][19] - The MSL team includes several researchers from OpenAI, with a notable representation of Chinese talent, indicating a strong focus on advanced AI research [24][27] Group 3 - Despite Zhao's appointment, Yann LeCun, a Turing Award winner, will continue as the Chief Scientist of FAIR, which focuses on long-term AI research [10][11] - The MSL aims to push the frontiers of superintelligent research, with a commitment to aligning AI with human needs [8][28] - The high percentage of Chinese members in the MSL team has led to discussions about the efficiency of communication within the team [25][27]

2025年7月22 — 24日，以 " 具身智能与多模态交互技术的融合与突破 " 为主题的2025智能机器人关键 技术大会在齐齐哈尔隆重举行。 本次大会由《机器人技术与应用》杂志社主办，山东大学、北京科技大 学、浙江工业大学、天津理工大学、齐齐哈尔大学和哈尔滨工业大学机电学院联合承办， 大会得到了 中 国自动化学会机器人专业委员会、中国人工智能学会智能机器人专业委员会、中国仪器仪表学会智能车 与机器人专业委员会和中国工程建设焊接协会机器人及智能焊接专业委员会 大力 支持， 汇聚了众多机 器人领域的顶尖专家、学者及行业精英 ，共同探讨具身智能领域前沿技术以及未来发展趋势。 大会同 期，蓝点触控（北京）科技有限公司 、北京诺亦腾科技有限公司、上海念通智能科技有限公司、烟台睿 感物联技术有限公司等企业携主打产品惊艳亮相并在现场演示，引不少观众驻足。 北方科技信息研究所韩志强 副 所长 、北方华安工业集团有限公司董事程振轩、 齐齐哈尔大学王志刚 副 校长 、 国家科技部二级研究员刘进长、山东大学李贻斌 教授 、工信部赛迪研究院科技处董凯 处长 、 哈尔滨工业大学付宜利 教授 、清华大学刘辛军 教授 、西安电子科技大学 ...

Core Viewpoint - Cloud Voice, a general artificial intelligence "unicorn," successfully listed on the Hong Kong Stock Exchange on June 30, after a tumultuous journey to the capital market, including a failed IPO attempt on the Shanghai Stock Exchange and two unsuccessful applications for the Hong Kong listing [1] Group 1: Company Overview - Cloud Voice issued 1,560,980 shares globally, with the Hong Kong public offering being oversubscribed by 91.66 times, leading to an adjusted sale of 624,400 shares [1] - The company set its IPO price at HKD 205 per share, raising approximately HKD 320 million, ranking 31st among 43 companies newly listed on the Hong Kong Stock Exchange in 2025 [1] - The company’s stock price surged on its debut, reaching a peak of HKD 319.80, a 56% increase, and closing at HKD 296.40, a 44.6% rise from the IPO price [2] Group 2: Investment and Financing - Major cornerstone investors included SenseTime, Runjian Co., and Zhenyi Asset Management, collectively subscribing to 462,860 shares, accounting for nearly 30% of the offering [2] - Prior to the IPO, Cloud Voice completed 10 rounds of equity financing, attracting investments from various institutions totaling approximately CNY 2.436 billion [2] - The company’s valuation reached approximately CNY 8.333 billion (USD 1.929 billion) after the completion of the D3 round of financing in May 2023 [2] Group 3: Financial Performance - The company reported cumulative losses of CNY 1.205 billion from 2022 to 2024, with a projected single-year loss of CNY 454 million in 2024, marking a 21.4% year-on-year increase [8] - Operating cash flow has been negative for three consecutive years, with trade receivables significantly high and turnover days exceeding industry averages [8] - The customer retention rate in the medical services sector dropped from 70.4% in 2022 to 53.3% in 2024, indicating challenges in maintaining client relationships [8] Group 4: Market Position - As of 2024, Cloud Voice ranked fourth in the Chinese AI solutions provider market, with a market share of only 0.6%, highlighting a significant gap compared to the top three competitors [8]

Core Insights - The discussion revolves around the potential of AI to reach a transformative point akin to AlphaGo's "move 37," suggesting that AI may be approaching a critical technological shift [1][30] - Demis Hassabis warns of the risks associated with AI advancements, emphasizing the need for cautious optimism [1][30] Group 1: AI and Natural Systems - Hassabis believes that all natural models can be efficiently modeled through classical learning algorithms, particularly in fields like biology, chemistry, and physics [4][5] - The probability of achieving Artificial General Intelligence (AGI) by 2030 is estimated at around 50%, with benchmarks including the ability to propose new scientific hypotheses [4][30] - AI systems like AlphaGo and AlphaFold demonstrate the capability to solve complex problems through intelligent guided searches [4][5] Group 2: AI's Understanding of Reality - The Veo 3 model showcases an impressive ability to generate realistic videos and demonstrates a form of "intuitive physics" understanding [7][8] - Hassabis expresses surprise at Veo 3's ability to learn from video observation without physical interaction, challenging previous assumptions about AI's need for embodiment to understand the physical world [9][10] Group 3: Future of Gaming with AI - Future gaming experiences may be revolutionized by AI, allowing for dynamic story generation based on player decisions, creating a more immersive experience [12][13] - Hassabis envisions a future where AI can create truly open-world games that respond in real-time to player choices, enhancing the gaming experience [12][13] Group 4: Evolutionary Algorithms and AI Innovation - The recently released AlphaEvolve system utilizes evolutionary algorithms to explore new solution spaces, combining large language models with evolutionary computation [18][19] - Hassabis believes that understanding the underlying dynamics of systems is crucial for discovering new solutions and that evolutionary computation can lead to significant breakthroughs [18][19] Group 5: AI's Role in Scientific Research - Hassabis discusses the concept of "research taste," suggesting that while AI can solve complex problems, it currently lacks the ability to propose profound scientific hypotheses [22][23] - The challenge lies in AI's ability to discern the right questions and hypotheses, which is a critical aspect of scientific research [23][24] Group 6: Future Energy Sources - Hassabis predicts that nuclear fusion and solar energy will become the primary energy sources in the future, addressing energy challenges and potentially leading to a Kardashev Type I civilization [43][44] - The development of efficient solar materials and nuclear reactors could enable humanity to harness abundant, clean energy [43][44] Group 7: Competition in AI Development - Hassabis emphasizes the importance of collaboration in AI research, stating that the goal is to safely bring technology to the world for the benefit of humanity [47][48] - The competition for talent in AI is intensifying, with companies like Meta employing aggressive strategies to attract top researchers [51]