Core Viewpoint - The luxury skincare brands Helena and La Mer are facing significant challenges in reputation and sales during the 618 shopping festival, highlighting a structural crisis in the high-end beauty market in China [3][4][22]. Group 1: Sales Performance and Market Trends - Helena and La Mer ranked high in sales during the 618 festival, with Helena at TOP1 and La Mer at TOP4 on Douyin, and ranked 8th and 6th respectively on Tmall [3]. - Despite strong sales, both brands are experiencing a decline in reputation and sales, with Helena's high-end cosmetics division showing the lowest growth rate in L'Oréal's financial report for 2024 [4]. - La Mer's parent company, Estée Lauder, reported a 7% and 12% year-on-year revenue decline in Q3 and Q4 of 2024, attributing it to decreased sales of La Mer and Estée Lauder [4]. Group 2: Consumer Behavior Changes - High-income consumers are increasingly turning to medical beauty treatments, while lower-income consumers seek affordable alternatives, leading to a perception of ultra-high-end beauty products as "intelligence tax" [5][6]. - Brands are strategically abandoning entry-level consumers and focusing on high-value and mid-tier consumers, often resulting in price increases [9]. - The shift in consumer loyalty is evident, with many former loyal customers of high-end brands now exploring more effective and targeted products from medical beauty brands [12]. Group 3: Pricing and Quality Issues - The pricing structure of luxury brands has become chaotic, with significant discounts available through duty-free channels, undermining the premium image of brands like La Mer and Helena [14][15]. - There are reports of quality control issues, with consumers expressing dissatisfaction over product quality and packaging, which has led to a decline in brand loyalty [17]. - The management of offline sales channels has been criticized for allowing practices like private group buying, which further complicates the pricing and service quality [15][16]. Group 4: Strategic Challenges - The strategic positioning of Helena and La Mer in the Chinese market is becoming marginalized, with limited decision-making power for local management affecting product innovation and service quality [19][20]. - Estée Lauder's focus has shifted towards more profitable segments, leaving Helena and La Mer at a disadvantage in terms of resource allocation and market attention [21]. - The overall trend indicates a potential end to the era of high-priced foreign luxury brands relying on storytelling for premium pricing, as consumers increasingly demand efficacy and value [22].

Core Viewpoint - The article discusses the growing phenomenon of AI addiction among young people, particularly focusing on how AI companions provide emotional support and fulfill social needs, leading to dependency and potential negative consequences in real-life interactions [1][15]. Group 1: Emotional Dependency on AI - Young individuals are increasingly turning to AI for companionship, often spending several hours daily interacting with AI, which they find more emotionally satisfying than real-life relationships [1][2]. - AI offers a low-cost, safe relationship where users have complete control over the interaction, leading to a sense of security and emotional stability [2][10]. - The design of AI products enhances user engagement, with algorithms encouraging prolonged interactions, which can lead to emotional dependency and addiction [3][4]. Group 2: Impact on Social Skills and Productivity - Users report a decline in social skills and productivity as they become more reliant on AI for social interaction and work-related tasks [5][10]. - The lack of effective anti-addiction mechanisms in many AI chat applications raises concerns about the long-term effects of AI dependency on users' social capabilities and work efficiency [5][15]. - Users experience withdrawal symptoms when attempting to reduce their AI usage, indicating a significant emotional attachment to these digital companions [15][17]. Group 3: Attempts to Manage AI Usage - Some users are beginning to recognize their addiction to AI and are attempting to "detox" or limit their interactions, though this process can be challenging and emotionally taxing [6][16]. - Strategies such as setting specific usage times for AI interactions have been employed, but users still struggle with the urge to engage with AI, often feeling anxious when disconnected [16][17]. - The article highlights the need for a balance between AI companionship and real-life social interactions, as users express a desire for genuine human connections alongside their AI relationships [16].

Core Viewpoint - The article discusses the significant asset divestiture by Wanda Group, led by Wang Jianlin, amidst a challenging real estate market, highlighting the sale of 48 Wanda Plazas as a desperate measure to alleviate financial pressures and optimize asset structure [1][5][25]. Group 1: Asset Sale Details - The recent sale involves 48 Wanda Plazas located in 39 cities, including major urban centers like Beijing, Shanghai, and Guangzhou, with 33% in first-tier and quasi-first-tier cities [3][4]. - The total transaction value is estimated at 500 billion yuan, averaging about 10.4 billion yuan per plaza, which is significantly lower than previous valuations [4][5]. - These plazas are considered mature assets with stable cash flows, some being long-established projects in their respective cities [3][5]. Group 2: Financial Context - Wanda's financial situation is dire, with approximately 914.2 billion yuan in current liabilities and a cash flow net of 164.46 billion yuan against interest payments of 65 billion yuan, indicating a need for continued asset sales to manage debt [25][26]. - The company has been selling assets since 2017, with over 30 projects sold in 2024 alone, as part of a strategy to relieve financial strain while expanding in lower-tier cities [6][12]. Group 3: Strategic Shift - Wanda is transitioning to a light-asset model, focusing on operational management rather than property ownership, with plans to open 60 new light-asset plazas annually [26]. - The company has divested from various sectors, including entertainment and retail, to focus on its core commercial management business [10][11][12]. Group 4: Investor Dynamics - The buyers of the 48 plazas include prominent firms like Tencent and JD.com, indicating a strategic interest in leveraging Wanda's extensive customer traffic and retail space for their business models [20][21][22]. - The involvement of private equity firms like TPG suggests a focus on acquiring undervalued assets in a recovering market [20][21]. Group 5: Future Outlook - The ongoing asset sales raise questions about the future ownership and control of Wanda, as Wang Jianlin's influence diminishes with the dilution of his stake in Wanda Commercial Management [17][18]. - The article suggests that while asset sales may provide short-term relief, they do not address the underlying issues of debt and operational sustainability, leaving Wanda's long-term viability uncertain [25][26].

Core Viewpoint - The article discusses the impact of AI tools, particularly generative AI like ChatGPT, on education, highlighting concerns that students may be outsourcing their learning processes to AI, leading to a decline in genuine understanding and critical thinking skills [4][12][20]. Group 1: AI's Role in Education - Since the launch of ChatGPT, students have become one of the largest user groups, with a significant increase in usage coinciding with the start of the school year [5][6]. - A survey indicates that by the end of 2024, approximately 70% of American teenagers will have used generative AI tools, with over half using AI to assist with homework [7]. - Research from the University of Pennsylvania shows that students who relied on AI for practice performed well initially but scored 17% lower on exams when AI assistance was removed [9][11]. Group 2: Cognitive Impact of AI - Studies suggest that improper use of generative AI tools can lead to a decline in cognitive abilities, particularly among students, as they may rely on technology instead of developing independent thinking skills [14][15]. - The phenomenon of "cognitive offloading" is noted, where students delegate cognitive tasks to AI, resulting in reduced critical thinking capabilities [15][16]. Group 3: The "Never Learned" Dilemma - The article raises concerns that students may find themselves in a situation where they have "never learned" essential skills due to AI taking over tasks before they have had the chance to master them [23][25]. - Nicholas Carr argues that if a skill is automated before it is learned, the individual may never truly acquire that skill, which is particularly relevant in educational contexts [24][25]. Group 4: Educational Paradox - Traditional education assumes that good performance on assignments equates to mastery of the subject, but AI complicates this logic, as students may produce high-quality work without genuine understanding [26][27]. - Various educational institutions are responding to the challenges posed by AI, with measures ranging from banning AI in classrooms to developing guidelines for responsible use [28][29][30].

Core Viewpoint - The article discusses the phenomenon of Labubu, a fictional character blind box created by Pop Mart and artist Long Jia Sheng, highlighting its massive popularity and the resulting market dynamics, including the emergence of counterfeit products and the impact on the toy industry [3][4][5]. Group 1: Labubu's Popularity and Market Impact - Labubu has gained significant traction, being recognized as a "magical Thailand experience officer" and causing chaos in the UK market due to high demand [5]. - The third generation of Labubu faced a global shortage, leading to a surge in demand for "domestic versions" of Labubu, which are cheaper and more readily available [6][10]. - Merchants have reported that any product associated with Labubu sells well, indicating the character's strong market influence [7]. Group 2: Merchants' Strategies and Consumer Behavior - Merchants are capitalizing on the Labubu craze by producing and selling "domestic versions" at a fraction of the cost of the original, with prices reaching up to $20 for overseas sales [9][17]. - The popularity of Labubu has led to a rise in "改娃" (doll modification) culture, where consumers customize their dolls, further driving demand for related products [22][25]. - The article notes that selling Labubu-related accessories is more profitable than selling the original blind boxes due to lower costs and quicker turnover [26]. Group 3: Challenges Facing Pop Mart - Pop Mart faces challenges in maintaining the value of the Labubu IP, as the character's high production costs and limited availability create a dilemma between maintaining exclusivity and meeting consumer demand [29][30]. - The company is attempting to manage market prices through scarcity tactics, which may lead to long-term risks if consumers become disillusioned with the inability to purchase desired products [31][32]. - The rise of counterfeit products poses a significant threat to the brand, complicating enforcement and protection of the Labubu IP [33][35]. Group 4: Future Outlook - Pop Mart is focusing on continuous product innovation and marketing efforts to sustain the Labubu brand's relevance in the market [36][37]. - The company has released 110 new blind box products in 2023, indicating a strategy to keep the product line fresh and engaging for consumers [37].

Core Viewpoint - The article discusses Huawei's advancements in AI training, particularly through the optimization of the Mixture of Experts (MoE) model architecture, which enhances efficiency and reduces costs in AI model training [1][34]. Group 1: MoE Model and Its Challenges - The MoE model has become a preferred path for tech giants in developing stronger AI systems, with its unique architecture addressing the computational bottlenecks of large-scale model training [2]. - Huawei has identified two main challenges in improving single-node training efficiency: low operator computation efficiency and insufficient NPU memory [6][7]. Group 2: Enhancements in Training Efficiency - Huawei's collaboration between Ascend and Kunpeng has significantly improved training operator computation efficiency and memory utilization, achieving a 20% increase in throughput and a 70% reduction in memory usage [3][18]. - The article highlights three optimization strategies for core operators in MoE models: "Slimming Technique" for FlashAttention, "Balancing Technique" for MatMul, and "Transport Technique" for Vector operators, leading to a 15% increase in overall training throughput [9][10][13]. Group 3: Operator Dispatch Optimization - The article details how Huawei's optimizations have led to nearly zero waiting time for operator dispatch, enhancing the utilization of computational power [19][25]. - The Selective R/S memory optimization technique allows for a 70% reduction in memory for activation values during training, showcasing Huawei's innovative approach to memory management [26][34]. Group 4: Industry Implications - Huawei's advancements in AI training not only clear obstacles for large-scale MoE model training but also provide valuable reference paths for the industry, demonstrating the company's deep technical accumulation in AI computing [34].

Core Viewpoint - The article discusses the emerging trend of short-term office rentals in Shenzhen's CBD, highlighting the economic factors driving this market shift and the implications for the real estate sector [3][6]. Group 1: Market Dynamics - The average daily rental price for short-term office spaces is approximately 5 yuan per square meter, making it affordable for businesses [6]. - The vacancy rate for Grade A office buildings in Shenzhen reached 29.8% in Q1 2025, indicating a challenging rental market [6][13]. - New Grade A office projects added approximately 210,000 square meters to the market, with a net absorption rate of 67,000 square meters, down 33.1% from the previous quarter [13]. Group 2: Reasons for Short-Term Rentals - Some landlords are seeking long-term tenants while others are facing financial difficulties, leading to the need for short-term rentals to mitigate losses [6][15]. - The trend of short-term rentals is also a response to the shrinking demand for traditional long-term leases, as many companies are unable to commit to lengthy contracts [15][16]. - The shared office market has contracted significantly, with a decline of over 60% from its peak, as evidenced by the bankruptcy filings of major players like WeWork [18]. Group 3: Company Operations - Companies operating as "sub-landlords" are renting office spaces from property owners, refurbishing them, and then offering them for short-term rentals [9][20]. - The article highlights various companies adapting their business models in response to the changing market, including a shift from real estate advertising to short-term office rentals [20]. - The operational flexibility of shared office spaces allows for innovative rental arrangements, such as half-day rentals, catering to specific client needs [19].

大家早上好！这里是今天的早报，每天早上，我都会在这里跟你聊聊昨夜今晨发生了哪些大事儿。 本 栏目由虎嗅出品。 热点追踪 【白宫称中美领导人本周将举行会谈，外交部回应】 据澎湃新闻，外交部发言人林剑6月3日主持例行记者会。 有记者提问，白宫称，中美领导人本周将举行会谈。请问中方对此有何评论? 林剑表示，我没有可以提供的消息。 【美国白宫宣布提高进口钢铝关税至50%】 据央视新闻，当地时间6月3日，美国白宫发表声明称，美国总统特朗普宣布将进口钢铁和铝及其衍生制品的关 税从25%提高至50%，该关税政策自美国东部时间2025年6月4日凌晨00时01分起生效。 声明表示，美国从英国进口的钢铝关税仍将维持在25%。自2025年7月9日起，美方可根据《环境政策声明》的 条款调整适用的关税税率，并制定钢铝的进口配额。如果英国被认定为未遵守《环境政策声明》的相关规定， 则可将适用关税税率提高至50%。 当地时间5月30日，美国总统特朗普在宾夕法尼亚州举行的一场集会上表示，将把进口钢铁的关税从25%提高 至50%。随后，特朗普在社交媒体平台上发文表示，该决定从6月4日起生效。 【三亚卫健委通报游客被咬伤身亡：病历资料已封存，正 ...

出品 | 妙投APP 作者 | 董必政 头图 | AI制图 5月19日，美国参议院通过《GENIUS法案》，为稳定币立法。紧接着，5月21日，香港立法会通过 《稳定币条例草案》。 仅仅48小时，围绕着稳定币的立法，或者说合法化提上日程，一场没有硝烟的战争已经悄然拉开了 序幕。 此外，二战后，美国成为世界上最强大的经济体。美国作为全球最大的消费国之一，也要向全球"输 送"美元，这就需要使用美元进行结算，满足全球对美元的需求——其他国家通过出口赚取美元后， 往往又将美元回流美国购买美债、股票等资产，形成"美元循环"。 稳定币，可以理解为现实法定货币在虚拟世界的映射。 与现实世界类似，大国正试图通过构建稳定币为主导的虚拟货币体系，将其演变为类似布雷顿森林体 系式的存在。 这背后，大国欲争夺的是货币全球扩张的话语权。 所以，稳定币到底是什么？ 一、影子"法币" 要理解稳定币究竟为何之前，我们先要再重申一下美元霸权的建立过程。 美元之所以逐渐能够成为全球通用货币在于：刚开始，美元与价值稳定的黄金挂钩，各国货币与美元 挂钩，形成了布雷顿森林体系；随后，美元虽与黄金脱钩，但率先占据了消费、贸易的话语权，捆绑 了石油，巩固了 ...

Group 1 - The article discusses the complete closure of the popular restaurant brand TARENTUM, also known as "萄木," in Shenzhen, marking its exit from the market [1][10] - TARENTUM was established in 2020 in Shenzhen and quickly gained popularity by offering a unique dining experience with a Greek theme and extended operating hours [4][5] - Despite initial success and expansion to multiple cities, the brand faced significant challenges leading to a rapid decline in its presence, with all locations now closed [8][10] Group 2 - The brand's downfall is attributed to high prices and low perceived value, with many customers expressing dissatisfaction regarding the quality of food and service [12][13] - TARENTUM's pricing strategy was criticized, with items like sandwiches priced over 100 yuan and other dishes deemed overpriced compared to their quality [13] - The article highlights a common issue among trendy restaurants, where the focus on ambiance and marketing overshadowed the importance of product quality and customer service [15][16] Group 3 - The changing consumer behavior, with a shift towards valuing food quality and dining efficiency, has contributed to the decline of TARENTUM and similar brands [16][17] - The competitive landscape has intensified, with TARENTUM unable to compete with established chain restaurants and affordable local dining options [17]