Core Viewpoint - Scientists have discovered a new approach to combat obesity by enhancing the body's own heat production through a novel experimental drug that stimulates mitochondria to burn energy more efficiently as heat rather than storing it [1][4]. Group 1: Research Findings - The research, led by a team from the University of Technology Sydney, focuses on a natural process known as "mitochondrial uncoupling," which reduces the efficiency of energy production in cells, leading to more energy being released as heat [4]. - The team has successfully designed a new generation of "mild" mitochondrial uncouplers that can control the heat production process within safe limits, avoiding the toxic risks associated with traditional drugs [4]. - This mild intervention may also help reduce cellular oxidative stress, potentially improving overall metabolic health and offering additional benefits such as delaying aging and related neurodegenerative diseases [4]. Group 2: Implications for Obesity Treatment - Current mainstream weight loss drugs often require injections and may have side effects, prompting the scientific community to seek safer strategies for promoting metabolic health [4]. - The breakthrough in this research provides a critical blueprint for designing a new generation of drugs that can safely induce mild mitochondrial uncoupling, which could lead to more effective obesity treatments in the future [4][5].

来源：光明日报 本报北京12月29日电（记者崔兴毅） 如何在不损伤文物的情况下，看清其细微的结构变化与老化痕 迹，一直是文物保护工作面临的巨大挑战。近日，中国科学院空天信息创新研究院（空天院）广州园区 研究员王振友团队研制的显微时间门控拉曼光谱仪，在三星堆象牙考古研究方面取得进展，成功为三星 堆象牙完成高清晰度的"分子级无损检测"，首次清晰揭示其跨越三千年的老化过程。 传统用于分析文物成分的拉曼光谱技术，在面对三星堆象牙时常常失效，这是因为长期埋藏使象牙产生 强烈的荧光，就像在文物表面打了一盏刺眼的灯，完全"淹没"承载物质信息的微弱拉曼信号，导致检测 无法进行。 研究团队的核心创新在于"时间门控"技术。他们发现，虽然拉曼信号和荧光信号同时发出，但拉曼信 号"闪现"极快（低于皮秒级），而荧光信号"滞留"时间较长（纳秒级）。新技术就像一个拥有超级快门 的探测器，只在拉曼信号闪现的瞬间打开进行捕捉，随即关闭，从而巧妙避开持续存在的荧光干扰。 王振友介绍，在对4块三星堆象牙碎片的研究中，2块用传统方法几乎"看不见"的碎片，在新技术的检测 下清晰呈现。其有效信号的信噪比提升了20倍以上。分析结果直观显示出象牙内部有机质 ...

Core Insights - 2025 is viewed as a pivotal year for AI transitioning from an "elite tool" to "widely available," reshaping human language practices, cognitive structures, and value judgments [1] - The mainstream discussion around AI is often focused on efficiency, cost, and competitive advantage, but its most profound impact extends beyond economic and industrial structures to the reconfiguration of human experience [1] Group 1: Impact on Humanities - Generative AI is challenging the traditional relationship between language and human subjectivity, as it can now perform complex language production, affecting how humans learn, express, and think [2] - The humanities are experiencing a defensive position, with concerns about being replaced by AI while simultaneously striving to prove their irreplaceability [2] - AI's ability to efficiently conduct literature reviews and style mimicry exposes the over-technicalization and formalization of certain humanities practices, prompting a reflection on whether the focus has shifted from studying "humans" to merely repeating discourse mechanisms [2] Group 2: Structural Crisis in Humanities - AI accelerates the revelation of existing divisions within the humanities, which were previously obscured by institutional evaluation systems and habitual discourse [2] - Research that begins with problem awareness, historical depth, and value judgment remains difficult for algorithms to replace, emphasizing the importance of understanding specific contexts and responding normatively to real-world issues [3] - The hollowing out of humanities practices that rely heavily on established paradigms and templates becomes evident as AI can easily replicate their operations [3] Group 3: Reevaluation of Humanities - The humanities must shift from "knowledge production" to "meaning judgment," from "text-centered" to "experience, history, and ethics-centered," and from "method-first" to "problem-awareness first" [4] - AI can generate language but cannot bear the responsibilities that come with it, nor can it truly experience human emotions or confront ultimate issues like death and justice [4] - The true test for the humanities in the age of AI is not whether they will be replaced, but whether they can still address fundamental questions about humanity [4]

Core Insights - The Arctic region has shifted from a "cooperative Arctic" to a "fragmented Arctic," influenced by both climate change and geopolitical tensions [1][2] Group 1: Climate Change Impact - The Arctic is warming at a rate four times faster than the global average, leading to accelerated melting and fragmentation of sea ice [1] - The natural fragmentation of the Arctic is characterized by the rapid loss of sea ice, which poses significant environmental challenges [1] Group 2: Geopolitical Tensions - The Russia-Ukraine conflict in February 2022 marked a turning point, leading to a division in Arctic governance and cooperation [2] - The Arctic Council, which has historically facilitated international cooperation, is facing its greatest challenge since its inception, resulting in a decline in collaborative efforts [2] - The geopolitical landscape is increasingly polarized, with Russia on one side and the Arctic Seven (U.S., Canada, and five Nordic countries) on the other [2] Group 3: Security Concerns - The absence of Russia from international Arctic security meetings is increasing, with upcoming conferences focusing solely on security issues [3] - The themes of recent Arctic security conferences reflect the growing division, with discussions centered on managing polarization and potential military confrontations [3] - The geopolitical significance of the Svalbard archipelago is rising, with concerns about military use challenging existing treaties [3] Group 4: Political Influence on Scientific Research - Political pressures are distorting the relationship between science and politics, leading to a climate of distrust among Arctic researchers [4] - The participation of U.S. representatives in Arctic conferences has drastically decreased due to security concerns and funding cuts, highlighting the impact of political dynamics on scientific collaboration [4] - The evolving narrative of the Arctic reflects a broader question about the role of human agency in shaping cooperation and fragmentation in the region [4]

| 三M MAU | 一 可 | 24 | 火 产品正式开始研发 2025年 3月12日 近日，AI原生应用企业与爱为舞对外发布首个真人级AI"一对一"导师落地近一年"成绩单"。数据显示，真人级AI导师教出的首批AI学习"原住民"规模已达百 万级，学员持续学习达316天，单个学员最长学习记录达9000分钟，单次课最高互动达274次。 公开信息显示，与爱为舞成立于2023年5月，核心创始团队深耕AI技术与教育场景多年。今年2月份，公司发布旗下教育品牌"爱学"，是国内第一款大规模落 地的真人级AI"一对一"导师产品，面向全年龄段人群提供可持续陪伴学习服务。 来源：环球网 爱学App正式上线 AI导师Lucy开启服务 2025年 10月24日 终身学习体系上线 AI导师团集结 致力破解"不可能三角"，首个真人级AI导师交出亮眼成绩单 教育行业长期面临"大规模-高质量-低成本"不可能三角的约束。大班课虽然成本可控，却难以照顾到每个学生的个体差异；线下一对一教学针对性强，但难 以规模化推广，成本也更高。生成式AI技术的突破，为攻克这一难题带来了新可能。 "生成式AI出现后，有可能让服务业变成制造业，让每一个人都被立即关 ...

来源：环球网 【环球网科技综合报道】12月29日消息，微信官方发布文章回应关于应用占用存储空间的误解，表示使 用超过40GB存储空间的用户中，聊天记录平均占比达到70%。 微信官方表示，基于用户需求开发出来的很多功能，不一定都会导致存储空间占用变多。例如小程序、 朋友圈、视频号等功能产生的数据都是"缓存"，是可再生的数据，在缓存中把这些数据删掉，不影响我 们继续使用这些功能，这些数据也随时可从服务器重新下载。同时，缓存有空间占用上限和自动清理机 制。 微信官方表示，对于很多被空间不足困扰的用户来说，问题根源可能是聊天记录。微信占用空间在 40GB以上的用户，聊天记录平均占比达70%。（思瀚） ...

来源：环球网 该公司于 2023 年首次宣布成立应急准备小组，称该小组将负责研究潜在的"灾难性风险"，无论是像网 络钓鱼攻击这样更直接的风险，还是像核威胁这样更具推测性的风险。 不到一年后，OpenAI 将应急准备主管调任至专注于人工智能推理的岗位。OpenAI 的其他安全高管也相 继离职或转任应急准备和安全以外的新职位。（思瀚） 【环球网科技综合报道】12月29日消息，据TechCrunch报道，OpenAI 正在招聘一位新的高管，负责研 究从计算机安全到心理健康等领域的新兴人工智能相关风险。该职位薪酬为55.5万美元外加股权。 OpenAI CEO奥尔特曼在一篇帖子中承认，人工智能模型"开始带来一些真正的挑战"，包括"模型对心理 健康的潜在影响"，以及"在计算机安全方面非常出色，以至于开始发现关键漏洞"的模型。 ...

Core Insights - Douyin's "Youth Creator Growth Plan" has successfully generated over 800 million yuan in cash income for 14,000 young creators within six months, with total views exceeding 5 trillion [1] - The plan has created new employment pathways for youth, with 260 creators joining the "Million Fans Club" and 1,590 surpassing 100,000 followers [1] Group 1: Youth Entrepreneurship and Support - Youth are increasingly engaging in entrepreneurship through various avenues, including e-commerce and content creation, driven by their innovative spirit and adaptability [1][2] - Government initiatives are in place to support youth entrepreneurship, including training programs and financial subsidies for skills development [2] - The rise of platform economies provides new opportunities for youth, allowing them to explore content creation as a viable career path [2] Group 2: Douyin's Role and Initiatives - Douyin collaborates with universities to offer courses on content creation, aiming to equip young creators with the necessary skills and knowledge [3] - The platform has implemented various support measures, such as traffic allocation and AI tool encouragement, enhancing content quality and creator engagement [4] - A significant percentage of creators on Douyin report substantial income, with many transitioning from personal expression to professional content production, contributing to job creation and industry growth [4] Group 3: Content Quality and Audience Engagement - Douyin's ecosystem has seen a 160% increase in quality content views and a 220% rise in viewing duration, driven by the contributions of young creators [4] - The platform fosters a balance between self-expression and value realization for creators, while also catering to diverse audience needs for quality content [5] - The ongoing demand for high-quality content remains strong, indicating a sustainable future for content entrepreneurship [5][6]

Core Insights - The collaboration between the Institute of Genetics and Developmental Biology and Alibaba Cloud aims to launch a Smart Breeding Program, which will create a standardized phenotypic data collection system and a digital brain for plant breeding [1][3]. Group 1: Collaboration Details - The Smart Breeding Program will utilize AI technology to automate the collection and analysis of vast biological omics data, transitioning from traditional experience-based breeding to intelligent design breeding [3][4]. - The partnership will leverage Alibaba Cloud's full-stack AI infrastructure to build a research support system encompassing hardware, cloud platforms, and the digital brain [3]. Group 2: Technological Implementation - A digital twin experimental platform for plant breeding will be co-developed to validate the decision-making effectiveness of the digital brain [3]. - The digital brain will serve as the core decision-making layer, utilizing extensive data from the Institute of Genetics and Developmental Biology to enhance breeding precision and efficiency [3]. Group 3: Strategic Importance - The collaboration is expected to significantly boost China's original innovation capabilities in biological breeding, providing robust technological support for national food security [4]. - The Qianwen platform has become a universal technical base for research, serving over 300 research institutions and universities, indicating its widespread acceptance and utility in the scientific community [4].

【环球网科技综合报道】12月29日消息，小米集团发布公告称，公司联合创始人、执行董事、副董事长 林斌计划自2026年12月起，在符合相关法律法规及香港联交所上市规则要求的前提下，每12个月出售金 额不超过5亿美元的公司B类普通股，累计减持总金额不超过20亿美元。 来源：环球网 公告显示，林斌此次减持计划所得款项主要用于成立投资基金公司。同时，林斌明确表达了对小米集团 业务前景的充分信心，并承诺将长期服务于集团。（思瀚） ...