来源：环球网 该报告基于SellCell内部以旧换新数据及40家独立iPhone收购商的二级市场价格趋势分析得出，iPhone 17 Pro Max在以旧换新Top20机型中占比达11.5%。其占比从去年11月末的5.1%增至今年2月初的11.5%， 约12周内实现翻倍，增长势头迅猛。值得注意的是，86%的被置换机型处于几乎全新或良好状态，可见 不少用户在购买后不久便选择置换。 【环球网科技综合报道】2月19日消息，据外媒macrumors报道，SellCell发布的最新以旧换新数据显 示，苹果最新旗舰机型iPhone 17 Pro Max上市仅数月，便迅速登顶独立以旧换新市场，成为当前最热门 的以旧换新智能手机，打破了旧机型主导以旧换新市场的常规格局。 对比来看，iPhone 15 Pro Max与iPhone 14 Pro Max表现依旧突出，各占以旧换新总量的7.3%；iPhone 16 Pro Max与iPhone 13各占7.2%。据悉，Top20机型合计占所有以旧换新活动的47%，而iPhone 17 Pro Max 作为新机能脱颖而出，实属罕见——通常以旧换新市场由保有量高的旧机型主导。 保值力 ...

Group 1 - The digital retail sector is experiencing significant changes, with various platforms facing challenges and opportunities in 2025 [1] - Douyin e-commerce is seeing an alarming daily return volume of millions of packages, indicating a shift in the dynamics between e-commerce and logistics industries [6] - Alibaba's anti-counterfeiting alliance reported a 58.5% decrease in counterfeit complaints on the Taotian platform compared to the previous year [8] Group 2 - The second-hand mobile phone recycling platforms are under scrutiny for practices like "high price estimation and low price upon receipt," raising user concerns about fairness and privacy [3] - The "old for new" service on Taobao reportedly saw a 32.7% reduction in the recycling price for an iPhone 15 Pro Max, highlighting issues of price manipulation [11] - A well-known women's clothing store with over 10 million followers announced its closure due to continuous losses, despite significant sales figures [12] Group 3 - Pinduoduo is tightening regulations on promotional language related to free gifts, requiring merchants to use official tools for compliance [15] - Pinduoduo has launched a special governance initiative to combat the use of pre-recorded videos masquerading as live streams, aiming to protect consumer rights [16] - Pinduoduo's new "delivery to village" service aims to enhance competitiveness in rural markets by facilitating logistics for merchants [17] Group 4 - Douyin e-commerce has introduced a new intelligent marketing product, "Qianchuan Chengfang," which integrates AI technology to optimize marketing strategies [30] - Douyin's nine support policies have reportedly saved merchants nearly 29 billion yuan in operating costs from January to November 2025 [32] - Douyin e-commerce's "treasure shop" report indicates a 165% year-on-year increase in GMV for participating merchants [29] Group 5 - The online retail sector in China saw a total online retail sales figure of 144.582 billion yuan from January to November 2025, reflecting a 9.1% year-on-year growth [67] - The Ministry of Commerce reported that online service consumption grew significantly, with a notable increase of 21.7% [77] - The National Development and Reform Commission has issued new rules prohibiting platforms from selling products below cost to prevent market disruption [68]

在将显示器熔化到金属外壳的过程中， LG生产的显示器未能通过可靠性测试。苹果正在调整显示屏的 设计，以便通使其过测试，而且该公司仍然可以使用三星的显示器进行组装。 这个制造问题或将对iPhone 15 Pro和Pro Max造成严重的冲击，导致许多消费者难以在第一时间购买到 新设备。 至少有一款iPhone机型出现供不应求的情况在此前也曾多次发生过，而且也曾出现过 iPhone机型全部售 完，人们要等待几个月才能看到苹果提高产量、满足需求的情况。在这种情况下，苹果可能的利润率可 能会受到一定的影响，其假日季度收入将低于预期。 据报道，苹果在iPhone 15 Pro和Pro Max的显示屏的生产上遇到了困难，这可能会导致九月份新机上市 的时候设备数量受限。 据悉，iPhone 15 Pro系列将比iPhone14 Pro机型的边框更薄，正是边框尺寸的减小造成了问题。 苹果的供应商正在使用一种新的显示器制造工艺来缩小面板尺寸，这给 LG Display公司带来了问题。苹 果在2019年也遇到过类似的问题，那就是Apple Watch Series 7的显示屏尺寸增加了，导致这款设备的上 市时间被延迟了一个月。 ...

Group 1: Public Fund Industry - The public fund industry in China, valued at over 36 trillion yuan, is undergoing significant reforms with the release of new guidelines for performance benchmarks [1] - The guidelines aim to standardize performance benchmarks and enhance investment management discipline, promoting stable investment styles and long-term returns for investors [1] Group 2: Li Auto - Li Auto announced a recall of 11,411 MEGA vehicles due to safety hazards linked to insufficient corrosion resistance in the cooling liquid, which could lead to battery thermal runaway [2] - The recall affects vehicles produced between February 18, 2024, and December 27, 2024, and is part of Li Auto's commitment to prevent future incidents [2] Group 3: Proya - Proya's Q3 2025 revenue was 1.736 billion yuan, a year-on-year decline of 11.63%, with net profit dropping 23.64% to 227 million yuan, marking the largest quarterly decline in recent years [3] - The main brand's revenue showed signs of stagnation, with a slight decrease in H1 2025, indicating challenges in maintaining growth [3] Group 4: Volkswagen - Volkswagen reported a net loss of 1.072 billion euros in Q3 2025, with a 61.5% year-on-year decline in net profit for the first three quarters [6] - Porsche's poor performance significantly impacted the group's overall results, with a 99% drop in operating profit to 40 million euros due to a shift in electric vehicle strategy [6] Group 5: ZTE Corporation - ZTE Corporation experienced an 88% drop in net profit in Q3 2025, with a revenue of 100.52 billion yuan, reflecting a 32.69% year-on-year decline in net profit [12] - The decline is attributed to reduced revenue from traditional communication services and pressure on margins in emerging computing businesses [12] Group 6: Eight Horse Tea - Eight Horse Tea's stock surged 86.7% on its debut, but the company faces challenges in a competitive market, holding only a 1.7% share in the high-end tea market [10][11] - Despite significant marketing investments totaling approximately 875 million yuan over the past three and a half years, revenue growth has been modest [11] Group 7: Xpeng Motors and WeRide - A public dispute arose between Xpeng Motors and WeRide over misleading data in Xpeng's roadshow materials, highlighting intense competition in the autonomous driving sector [7] - This incident occurred as both companies prepared for simultaneous listings, intensifying the rivalry in the market [7] Group 8: DJI - The chairman of Yingshi Innovation publicly criticized DJI for monopolistic practices, coinciding with DJI's significant market share in the panoramic camera sector [8] - DJI captured 43% of the global market share shortly after launching its first product, indicating fierce competition in the consumer electronics space [8] Group 9: Vanke - Vanke reported a net loss of 28.02 billion yuan in the first three quarters of 2025, with a revenue of 161.39 billion yuan and a low gross margin of 2.0% [9] - The losses are attributed to declining project settlement scales, low gross margins, and increased provisions for inventory depreciation [9]

Core Viewpoint - The article discusses a consumer's negative experience with a trade-in service provided by a company named "同城帮" (Tongcheng Bang), highlighting issues of price manipulation and lack of transparency in the evaluation process of used phones [1][5][19]. Group 1: Consumer Experience - A social media influencer reported a significant price drop in the trade-in value of their iPhone from an initial assessment of 4488 yuan to 3017 yuan after the phone was received, a decrease of 1471 yuan [1]. - The influencer questioned the validity of the price adjustment, citing a lack of visible damage on the phone as per the official assessment photos [4]. - Many other consumers have reported similar experiences of drastic price reductions during the trade-in process, indicating a pattern of potential price manipulation by the service provider [7][9]. Group 2: Company Background - "同城帮" is a service under Beijing Qili Software Technology Co., Ltd., which is part of the 360 second-hand goods ecosystem, providing comprehensive recycling and trade-in solutions [19]. - The company has raised over 200 million yuan in funding, with notable investors including Suning and 360 Group, indicating a significant backing in the circular economy sector [19][20]. Group 3: Customer Service Issues - The influencer attempted to escalate the issue through customer service, demanding evidence for the price reduction and expressing concerns over the lack of proper handling during the evaluation process [14]. - After multiple complaints, the company offered to increase the trade-in price from 3200 yuan to 4200 yuan, but the influencer refused, insisting on the return of the phone instead [15][16]. - There were indications of potential deceptive practices in customer service interactions, where the influencer was advised to cancel the return request, which could have led to losing the phone [18].

Core Viewpoint - Apple has demonstrated the feasibility of fine-tuning large language models (LLMs) on iPhones using a new method called Memory-Efficient Backpropagation (MeBP), which offers better trade-offs between memory usage and computation time compared to existing methods [1][4]. Summary by Sections Introduction - The article discusses Apple's recent paper on MeBP, which allows for model fine-tuning on resource-constrained mobile devices like the iPhone 15 Pro Max [1][3]. Methodology - MeBP focuses on using LoRA for fine-tuning LLMs, aiming to keep memory usage below 1GB, as recommended by PocketLLM [4]. - The fine-tuning process using MeBP consists of three main steps: compressing base model weights, implementing gradient checkpointing, and creating an efficient runtime for executing the training graph [5][10]. Model Weight Compression - The team employed 4-bit symmetric INT4 quantization for non-LoRA parameters, including embeddings, to reduce disk space usage [7][10]. Gradient Checkpointing - The LLM is divided into blocks to ensure that memory consumption during backpropagation remains within device limits. Automatic differentiation is used to generate a backward graph for each block [8][9]. Runtime Implementation - The MeBP runtime is designed to minimize memory usage by memory-mapping compressed model weights and only decompressing them on demand during training [15][16]. Experimental Performance - The team compared MeBP with MeZO, the only known optimization method for mobile LLM fine-tuning, using server-side simulations and performance evaluations on mobile devices [18][24]. - The experiments were conducted on models with parameters ranging from 0.5B to 4B, focusing on loss and next token accuracy as evaluation metrics [20]. Utility Comparison - Results indicated that while zero-order (ZO) optimization showed slower convergence compared to first-order (FO) optimization, MeBP significantly outperformed ZO in terms of convergence speed and computational efficiency [23]. Performance Comparison - MeBP was implemented in Swift on an iPhone 15 Pro Max with 8GB RAM, showing that MeBP's computation time per gradient step was 43% to 94% longer than MeZO, but it converged faster overall due to fewer required steps [24][28]. - The memory usage of MeBP was slightly higher than MeZO in the worst case, but overall training memory usage was approximately 10 times smaller than previous mobile implementations [28]. Conclusion - All tested LLMs could be efficiently fine-tuned within 1GB of memory, making them suitable for background training on mobile devices [28].

Group 1 - The article expresses a sense of frustration and longing for home, particularly the comfort of a mother's care and meals [6][8][7] - It highlights the challenges of work life, including stress and financial issues, leading to thoughts of quitting and returning home [6][7][8] - There is a recurring theme of seeking emotional support from family, especially mothers, during tough times at work [5][6][7] Group 2 - The article discusses the importance of work-life balance, emphasizing the need to separate work from personal life and enjoy leisure time [17][19] - It encourages improving efficiency during working hours to allow for more personal time, which can enhance creativity [19] - The message promotes the use of tools to optimize both work and life, suggesting a proactive approach to managing responsibilities [19]

Group 1 - The article discusses the emotional struggles of workers who often express a desire to resign, highlighting the importance of support from family during tough times [1] - It emphasizes that financial stability or a strong support system allows individuals to consider leaving their jobs without fear [1] - The piece reflects on the happiness that comes from receiving encouragement from family when expressing the wish to quit a job [1] Group 2 - The article includes humorous anecdotes and relatable content about the daily life of workers, showcasing their frustrations and desires for a better work-life balance [21][23] - It mentions the importance of separating work from personal life, suggesting that work should ideally be completed during designated hours [21] - The narrative captures the essence of modern work culture, where many individuals feel overwhelmed yet find solace in shared experiences and humor [40][67]

Core Viewpoint - Xiaomi has taken the incident of a car allegedly driving away on its own seriously, establishing a special task force to investigate and verify the situation with users [1][2]. Group 1: Incident Investigation - Xiaomi's spokesperson stated that the company formed a special task force to investigate the viral video of a car driving away on its own, collaborating with users to verify the facts [1]. - The investigation involved checking the vehicle's backend data and operation logs from the users' smartphones, concluding that the vehicle's data matched the commands issued from an iPhone 15 Pro Max, ruling out any quality issues with the vehicle [1]. - The vehicle's backend data indicated that it received a parking assist command from the iPhone 15 Pro Max during the reported time frame, which activated the parking assist feature and led to the vehicle exiting the parking space [1]. Group 2: User Communication and Apology - During user feedback and online customer service interactions, there was confusion regarding the device that issued the parking assist command, with customer service mistakenly identifying it as an iPhone 16 instead of the correct iPhone 15 Pro Max, leading to misunderstandings [2]. - Xiaomi expressed apologies for the confusion caused by the inaccurate information and committed to improving service processes in the future [2]. - The company emphasized that user safety and experience remain its top priorities and thanked users for their support [2].

Core Viewpoint - The incident involving a Xiaomi car reportedly driving away on its own has been investigated, and the company concluded that the vehicle's movement was triggered by a parking assist command sent from the user's phone, ruling out any quality issues with the vehicle [1][2]. Group 1 - Xiaomi has established a special task force to investigate the incident and has collaborated with the user to verify the situation [1]. - The investigation revealed that the vehicle received a parking assist command from the user's iPhone 15 Pro Max, which activated the parking assist feature, leading to the vehicle moving out of its parking space [2]. - Xiaomi acknowledged a mix-up in communication where customer service mistakenly identified the device that sent the command, which caused confusion for the user, and the company has apologized for this error [2].