跨 Consumer / Brand / Beauty / Luxury / Tech / Global 真实来源的结构化情报。Beauty 只是其一， 所有模块支持跨行业扩展。点击条目进入「知识卡」详情（原始情报 + 行业 / 营销 / 职业萃取）。
● 本地档案Circle K is more openly promoting and expanding its retail media network, Full Circle Media, as a way for advertisers to reach shoppers across its convenience stores, fuel pumps and digital channels in the U.S., Canada and Europe.
The conversation around AI has moved well beyond experimentation. At the recent Glossy & Modern Retail Brand Leaders Dinner, with partner Global Payments, founders and executives shared how they're using artificial intelligence to automate operations, rethink hiring, build internal knowledge platforms and refine marketing strategies. But despite the rapid adoption, attendees agreed that the businesses best positioned to win will still be those that invest in community, physical retail and experiences AI cannot replicate.
· MacBook Air 严重缺货，许多新订单延至 9 月交付 · 曾爱玲加入 B 站，负责 AI 视频生成业务 · Pixel 11 全系规格与售价曝光 #欢迎关注爱范儿官方微信公众号：爱范儿（微信号：ifanr），更多精彩内容第一时间为您奉上。
AI 产业扩张引发内存芯片短缺，消息称苹果 MacBook Air 笔记本出现供应紧张局面 8 月 3 日消息，据彭博社记者马克 · 古尔曼（Mark Gurman）报道，全球内存芯片短缺正在影响苹果 Mac 产品供应，其中受消费者欢迎的 MacBook Air 正面临库存压力。 古尔曼表示，此次内存短缺主要受到 AI 产业快速扩张影响。随着人工智能企业大量采购高性能存储芯片，挤压了消费电子市场供应，苹果旗下部分产品已经受到影响。 此前，用户基数更少的 Mac mini 和 Mac Studio 已经出现供应限制，而如今这一问题正在扩展至销量更高的 MacBook Air。目前，通过苹果官网购买 MacBook Air，部分型号预计需要等待至 8 月下旬才能发货，部分特殊配置甚至要等到 9 月。同时多家第三方零售商也声称，目前 MacBook Air 库存情况比以往更加紧张。
Circles uses the OpenAI API and Codex to power AI-native telco experiences, increasing ARPU by 22%, reducing churn by 9%, and improving development efficiency.
马斯克和 Altman 达成共识：人类已迈入 AI 奇点时代 科技媒体 thenews 近日发布博文，报道称马斯克和 Altman 在 AI 奇点（singularity）方面达成共识。 7 月 25 日作客 Relentless 播客节目时，Altman 表示：「我们现在差不多已经身处奇点时代。」奥尔特曼还预测，在人类目前能完成的工作中，AI 已经能够自动化完成约 30~40% 的工作，并表示 AI 正推动「自工业革命以来最大的经济变革」。 马斯克于 7 月 31 日在 X 平台发布推文，指出：「AI 在很多方面都已经超越了人类。我们正处于奇点之中。」 注：AI 奇点是未来学家雷 · 库兹韦尔（Ray Kurzweil）在 2005 年提出的假设，是一种理论情境，即人工智能自我改进的速度超过了我们人类监控和控制它的能力。(来源：IT 之家) 高通完成收购 Modular，加
DeepSeek-V4-Flash 正式版 API 上线公测，V4-Pro 正式版将「尽快」发布 7 月 31 日消息，今天（31 日）下午，DeepSeek 通过 API 文档发布日志，宣布 DeepSeek-V4-Flash 正式版 API 上线公测。 根据介绍，DeepSeek-V4-Flash 的 Agent 能力大幅增强，基准测试 远超 V4-Pro-Preview ： Terminal Bench 2.1: 82.7 NL2Repo: 54.2 Cybergym: 76.7 DeepSWE: 54.4 Toolathlon verified: 70.3 Agent Last Exam: 25.2 Automation Bench (Public): 25.1 DSBench-FullStack: 68.7 DSBench-Hard: 59.6 据悉，正式版 V4-Flash 原
OpenAI shares new results on long-standing open problems in mathematics and theoretical computer science, including advances in geometry, cryptography, and complexity.
OpenAI shares how its safety, security, transparency, and provenance practices support responsible AI governance in Europe. The work will continue as the EU AI Act advances.
A full-stack approach to making advanced AI more capable, more affordable, and more widely useful.
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Montana’s plan to become an experimental medical hub just pushed forward As of this week in Montana, biotech companies whose drugs have been through preliminary testing—sometimes in as few as 10…
Anomaly helped imagine the first day of school at its most musically surreal while building on the retailer’s brand platform.
作者｜Moonshot 编辑｜郑玄 最近，影视飓风开始进入 AI 影视赛道，选择与 TapNow 推出一套 AI 影像创作教程。 一套教程的出现，本身就是个信号，它说明 AI 影像正在走出「展示奇观」的阶段，开始形成可以被讲解、学习和复用的创作方法。 被教程吸引进来的，也未必都是准备拍电影的人。 车企、游戏公司和消费品牌都在增加视频内容的使用：产品发布需要概念片，游戏上线需要世界观短片，市场部门则要持续生产品牌视频、更新社交媒体。 对这些团队来说，AI 影像正在从一项新技术，变成工作里可以调用的制作能力。 但真正开始做视频以后，问题很快就从「怎么生成」变成了「怎么把视频做下去」。 一条完整的视频还要经过创意开发、视觉设定、分镜、素材生成、修改和剪辑。视频一旦从十秒左右的镜头走向完整作品，人物、场景和前后决定都要保持连续，原本分散在不同工具里的问题也会同时出现。 观众看到的是一条视频，但创作
越来越多的具身技术正先一步外溢到消费电子领域，包括这款会跳的机器人。 作者｜Li Yuan 编辑｜ 郑玄 会飞的无人机，我们见过。 会跳舞的人形机器人，我们也见过。 但你见过，行走方式纯粹靠跳的机器人吗？ 近期极客公园了解到，眺月科技正在做这样一款硬件——长得像无人机加了条腿。为了让世界认识跳跃机，他们年底将发布第一款遥控产品，瞄准的是穿越机发烧友和科技潮玩爱好者。 不管是机器狗还是人形机器人，腿脚再灵活，通常也迈不过比自己还高的坎。 这台机器不一样，原型机能跳到 2 米高，是自身高度的好几倍 ，人形、四足机器人过不去的障碍，它一跃而过，实在跨不过的沟壑，甚至可以直接飞过去，扛着它直接去野外玩也没问题。 这款硬件被定义成了一个新品类，他们给它起了个自己的名字，跳跃机，英文 hopcopter。 在它之前，已经有机器人能连续跳了——2016 年
See how Univé built an AI-ready workforce with ChatGPT Enterprise by combining leadership, responsible governance, and employee-led innovation to transform work at scale.
Celebrity stylist Micaela Erlanger explains on the Glossy Podcast why she prefers subtle “meta dressing” over literal red-carpet costumes, as Zendaya and Law Roach turn the “Spider-Man: Brand New Day” press tour into another fashion narrative.
Celebrity stylist Micaela Erlanger explains on the Glossy Podcast why she prefers subtle “meta dressing” over literal red-carpet costumes, as Zendaya and Law Roach turn the “Spider-Man: Brand New Day” press tour into another fashion narrative.
In this week's Luxury Briefing, I take a deep dive into the most significant group of earnings from recent months from LVMH, Kering, Hermés, Prada Group, Brunello Cucinelli and EssilorLuxottica – with analyst takes from Bernstein's Luca Solca. It seems like as brands are adjusting to slow growth, they're trying to meet customers at all price points, not always successfully. Also, how Frasers is buying up, or buying into more of the U.K. luxury space, and what Arnault said on X in reply to the Le Monde investigation.
OpenAI disrupted a Cambodia-based scam operation using ChatGPT to support investment, romance, gambling, and impersonation schemes.
As of this week in Montana, any biotech company with an experimental drug has a clear path to selling it to consumers. Companies whose drugs have been through preliminary testing—sometimes in as few as 10 healthy people—can pay $12,500 to apply to a newly established review board for approval. Once its treatment is rubber-stamped, the…
The app is focusing on partnerships to “meet our clients where they are,” said Vice President of Global Agency and Strategic Accounts Adrian Mulryan.
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. A fundamental flaw leaves LLMs strikingly vulnerable to attack It is impossible to make large language models fully secure against hacks because of a fundamental flaw in how they work, a…
Here’s the question most teams are actually trying to answer: do you need a standalone tool to monitor how your brand shows up in AI search, or do you need something that connects that monitoring to the content and CRM workflows that let you act on it?
It is impossible to make large language models fully secure against hacks because of a fundamental flaw in how they work, a team of researchers argue in a paper presented at the International Conference on Machine Learning, a top AI conference, this month. The claim has huge implications for the safety of this technology, which…
