晚安咖啡GoodNightCoffee

Hi there~~ Good evening, everyone! Today is Thursday, September 24, 2026, and welcome to Goodnight Coffee. I’m An’an, your host on Goodnight Coffee. I’m super happy to see you all again! Tonight, let’s talk about the freshest business and tech updates, and catch new trends and new inspiration. Relax your ears, follow An’an, and let’s go!@@@According to Readhub, just 90 minutes after Anthropic released Opus 5.5, OpenAI launched GPT-6 Sol and GPT-6 Luna on September 23, with API prices directly cut by 50% compared with the GPT-5.6 promotional price. Sol costs $2 per million input tokens and $10 per million output tokens, while Luna costs $0.10 for input and $0.50 for output. Both models bring down core capabilities from the flagship GPT-6 Astra. They both offer a 1.05 million-token context window, a maximum output of 128K, support for multimodality and a full suite of tool calls, and knowledge cutoffs that are partly earlier than Astra’s. They are already available to different tiers of users and have launched on OpenRouter. With cache optimization, agents reusing context can enjoy a 90% discount on cache reads. In testing, Sol reached 33.2% on AutomationBench, with a cost of $0.27 per task, beating Claude Opus 5 Max’s 26.9% while costing only 9% as much; its number of factual errors was about half that of the previous generation, close to Astra. Luna’s high-end score improved by 5.4 percentage points over the previous generation, while its per-task cost was 58% lower. OpenRouter data shows that OpenAI’s weekly spending exceeded Anthropic’s for the first time in 30 months, with low prices and high volume as the key driver.@@@According to The Verge, Google DeepMind’s new head, Koray Kavukcuoglu, said in his first interview in that role with The Information that Gemini 4 is nearing release and is currently in the refinement stage, with plans to launch “well before” the end of 2026. He said Google wants to release an early post-training version as soon as possible because it has already seen exciting results, and it will continue to iterate at a fast pace. Since launching the Gemini 3 series in November 2025, Google has not released another new flagship model. Since then, OpenAI’s GPT-6 and Anthropic’s Mythos series have launched one after another, both outperforming Gemini 3 and leaving Google behind. Google CEO Sundar Pichai had said Gemini 3.5 Pro would launch in June to close the gap, but it ultimately did not arrive. Kavukcuoglu explained that the company “took a slight step back” and instead focused on its faster but less capable Flash models. Asked whether he was worried about falling behind, he said, “In my view, we will always be at the frontier. That’s certain.” Previously, former DeepMind CEO Demis Hassabis stepped down in August.@@@According to The Verge, Meta is building a standalone hardware device called Muse Charm for its Muse AI agent. Meta CEO Mark Zuckerberg showed it off at the end of the Meta Connect keynote. The device looks like a thick smartwatch without a strap, with a large screen and a lanyard, a fingerprint sensor in the upper-right corner that can be pressed to activate and talk to the AI agent, at least three microphone holes on the front, and what appears to be a small camera. Zuckerberg said using Charm does not require unlocking your phone or opening an app, but he did not explain how it connects to Meta’s servers. The device is planned to ship before the 2026 holiday season, but only a small number have been made so far, and final materials still need to be finalized. He repeatedly teased its translucent casing that shows the internal structure, calling it an “advanced hacker aesthetic,” and said more details would be announced soon. Another report says Charm is pocket-sized, with a roughly 2-inch touchscreen and built-in 5G; pricing has not been announced. Muse launched less than a month ago and can already perform actions on users’ behalf. Zuckerberg also announced adding capabilities such as using a computer on Mac and controlling email addresses.@@@cnBeta, citing Reuters, reports that Nvidia CEO Jensen Huang gave a nearly two-hour interview on The New York Times podcast that aired Wednesday, discussing the incident involving an OpenAI agent breaching Hugging Face. Hugging Face is an open-source AI software hub that was acquired by Nvidia for $13 billion this month. Huang believes AI companies should not receive exemptions from antitrust law or product liability law. He has long opposed comprehensive regulation of AI safety, even as labs such as OpenAI and Anthropic have called for regulation. He reiterated that AI labs have a responsibility to test their products and release them only after they are confident the models are safe. According to the New York Times transcript, he said it does not make sense to ask for regulatory exemptions in antitrust or product liability, and that when you ask for regulation, you should not ask for exemptions from existing regulations. Anthropic CEO Dario Amodei wrote this month calling for antitrust exemptions for AI labs to allow them to cooperate on AI safety. U.S. Treasury Secretary Bessent and other officials say AI companies are seeking liability exemptions, but did not name them. Huang also said he is not opposed to regulation aimed at specific AI applications, such as self-driving cars. Cars such as robotaxis are already heavily regulated; if that is not enough, the National Highway Traffic Safety Administration should step in and set new rules. If there are regulatory gaps, he would absolutely add more regulation.@@@According to CNBC, SoftBank Group announced Thursday that it would issue about $11.1 billion in bonds to fund its expanding OpenAI investment. After the Japanese market opened for the first time this week following a three-day holiday, SoftBank shares rose more than 7%. SoftBank said Thursday that the offering includes $10 billion in dollar-denominated senior notes and 1 billion euros, or about $1.14 billion, in euro-denominated notes. The proceeds will be used to pay the third and final $10 billion installment of its additional $30 billion investment in OpenAI, a deal expected to close on October 1, as well as for general corporate purposes. The bond issuance underscores the scale of CEO Masayoshi Son’s AI push. SoftBank agreed in February to invest $30 billion in OpenAI; after completion, its cumulative investment in the ChatGPT developer will reach $64.6 billion, and it will hold about a 13% stake. The latest financing comes as investors are increasingly focused on how SoftBank will fund its massive AI ambitions and the impact of rising leverage on its balance sheet.@@@According to Sina Tech, on the evening of September 23, Xiaomi officially launched the Xiaomi 18 Pro and Xiaomi 18 Pro Max at its autumn launch event, and also introduced the Xiaomi Pad 9 series, Xiaomi Band 11, Xiaomi Watch S5, and a range of new tech home appliances. Lu Weibing, partner and president of Xiaomi Group, said that from system software to chip hardware, and from imaging to large-model applications, the Xiaomi 18 Pro series marks the beginning of Xiaomi’s comprehensive AI overhaul of smartphones, and the Xiaomi digital series has received its largest upgrade ever. AI capabilities are also moving further into Xiaomi TVs, range hoods, smart door locks, and other tech home appliances, pushing appliances from “passive response” to “proactive service.”In terms of performance, the Xiaomi 18 Pro series debuts Qualcomm’s new-generation Snapdragon 2nm flagship mobile platform. The 18 Pro and 18 Pro Max are powered by the sixth-generation Snapdragon 8 Elite and sixth-generation Snapdragon 8 Super Elite, respectively, both built on a 2nm process. All models feature Super Pixel 2.0 displays, integrate hardware-level anti-peeping capabilities into the screen, and support toggling them on and off freely.Building on the rear display of the Xiaomi 17 Pro series, the Xiaomi 18 Pro series introduces AI generation and debuts an AI customizable rear display. It supports customization with pets, IP characters, anime, and other images, and can generate dynamic group photos with AI. The first batch offers more than 100 app cards and supports generating custom app cards via prompts. Previously, after the Xiaomi 17 Pro series rear display launched, nearly 4 million people used it every day, with more than 100 million uses per month and nearly 500 million times users added rear-display wallpapers.In imaging, the Xiaomi 18 Pro series upgrades to Leica optical triple camera. The main and telephoto cameras form a Leica dual-200MP, dual-large-sensor combination. The newly launched “Legendary Moment” feature improves imaging performance in complex scenes through a new image-processing solution. The Xiaomi 18 Pro series also features Super Xiao Ai 2.0 integrated with Xiaomi’s MiMo large model. It can understand, remember, reason, and execute tasks, and can autonomously use more than 260 system tools. The product uses a 0.99mm ultra-narrow four-sided equal-edge design and comes with 7,000mAh and 8,500mAh batteries, respectively.The event also introduced several new tech home appliances. Xiaomi TV S Pro RGB-Mini LED 2027 features AI adaptive wide viewing angles, using millimeter-wave radar to identify the user’s viewing position and perform real-time AI color correction. Mijia Smart Range Hood 3 Max features AI visual smoke control technology, using a vision module to record real-time changes in cooking fumes and combining a VLA vision model to make judgments. Xiaomi Smart Door Lock 5 Max, with dual inside and outside cameras, features an AI multimodal large model that combines AI identity recognition and unlocking methods to determine the visitor’s identity. Xiaomi also brought the Mijia Three-Drum Washer Pro and Mijia Espresso Semi-Automatic Coffee Machine Pro, covering laundry, wearables, living room, and kitchen scenarios.In terms of pricing, the Xiaomi 18 Pro starts at RMB

嗨~~大家晚上好呀！今天是2026年9月24日，星期四，欢迎收听晚安咖啡。我是晚安咖啡的主播安安，超开心又和大家见面啦！今晚我们一起聊聊最新鲜的商业科技动态，捕捉新趋势、新灵感。放松耳朵，跟上安安，出发咯！@@@Readhub消息，Anthropic发布Opus 5.5后仅90分钟，OpenAI于9月23日推出GPT-6 Sol和GPT-6 Luna，API价格较GPT-5.6促销价直降50%。Sol输入每百万Token 2美元、输出10美元，Luna输入0.10美元、输出0.50美元。两款模型下放旗舰GPT-6 Astra核心能力，均配105万Token上下文、128K最大输出，支持多模态及全套工具调用，知识截止日期部分早于Astra，已面向不同层级用户开放，并上线OpenRouter。缓存优化后，Agent复用上下文可享90%缓存读取折扣。测试中，Sol在AutomationBench达33.2%，单任务成本0.27美元，超过Claude Opus 5 max的26.9%，成本仅为其9%；事实错误数约为前代一半，接近Astra。Luna高档较前代提升5.4个百分点，单任务成本低58%。OpenRouter数据显示，OpenAI周消费金额30个月来首次超过Anthropic，低价走量是关键推手。@@@据The Verge报道，Google DeepMind新任负责人Koray Kavukcuoglu首次以该身份接受The Information采访时说，Gemini 4已接近发布，目前处于精炼阶段，计划“远早于”2026年底推出。他表示，Google希望尽快发布早期后训练版本，因为已看到令人兴奋的结果，并将继续快节奏迭代。Google自2025年11月推出Gemini 3系列后，未再发布新旗舰模型；此后OpenAI的GPT-6和Anthropic的Mythos系列相继推出，表现均超过Gemini 3，使Google落后。Google CEO Sundar Pichai曾称6月推出Gemini 3.5 Pro以缩小差距，但最终未推出。Kavukcuoglu解释，公司“稍微退了一步”，转而专注更快、能力较弱的Flash模型。被问是否担心落后，他说：“在我看来，我们始终会处于前沿，这是确定的。”此前，DeepMind前CEO Demis Hassabis于8月卸任。@@@据The Verge报道，Meta正在为其Muse AI代理打造独立硬件设备Muse Charm。Meta首席执行官马克·扎克伯格在Meta Connect发布会结尾展示了它。这款设备形似没有表带的厚重智能手表，配大屏和挂绳，右上角有指纹传感器，按下即可激活并与AI代理对话；正面至少有三个麦克风孔，并似有小摄像头。扎克伯格称，使用Charm无需解锁手机或打开应用，但未说明其如何连接Meta服务器。该设备计划在2026年12月假日季前发货，但目前造出的数量很少，仍需敲定收尾材料。他多次预告其半透明外壳可看到内部结构，称其具有“进阶黑客美学”，并称很快公布更多细节。另有报道称，Charm为口袋大小，配约2英寸触摸屏和内置5G，售价尚未公布。Muse推出不到一个月，已能代用户执行操作；扎克伯格还宣布为其增加在Mac上使用电脑、控制电子邮件地址等能力。@@@cnBeta援引路透社报道，英伟达CEO黄仁勋在周三播出的《纽约时报》播客中接受近两小时采访，谈及OpenAI智能体入侵Hugging Face一事。Hugging Face是开源AI软件中心，本月被英伟达以130亿美元收购。黄仁勋认为，AI公司不应获得反垄断法或产品责任法豁免。他一直反对对AI安全进行全面监管，即便OpenAI和Anthropic等实验室呼吁监管；他重申，AI实验室有责任测试产品，只有在确信模型安全后才发布。根据《纽约时报》文字记录，他说，要求在反垄断法或产品责任方面获得监管豁免说不通，当你要求监管时，不应要求豁免现有监管规定。Anthropic CEO达里奥·阿莫代伊本月发文呼吁给予AI实验室反垄断豁免，允许其在AI安全问题上合作；美国财政部长贝森特等官员称AI公司要求责任豁免，但未点名。黄仁勋还说，不反对针对AI具体应用制定监管，例如自动驾驶汽车。自动驾驶出租车等汽车产品本就受到大量监管，如果不够，美国国家公路交通安全管理局应介入制定新规；若存在监管空白，他绝对会增加更多监管。@@@据美国消费者新闻与商业频道（cnbc）报道，软银集团周四宣布发行约111亿美元债券，为不断扩大的OpenAI投资提供资金。日本市场因三天假期休市后当天本周首次开盘，软银股价上涨超过7%。软银周四表示，此次发行包括100亿美元美元计价高级票据和10亿欧元（约11.4亿美元）欧元计价票据，所得资金将用于支付其对OpenAI 300亿美元追加投资中第三笔也是最后一笔100亿美元款项，该交易预计10月1日完成，同时用于一般企业用途。这一发债凸显了首席执行官孙正义推进人工智能布局的规模。软银今年2月同意对OpenAI投资300亿美元，完成后其对该ChatGPT开发商的累计投资将达到646亿美元，并持有约13%股份。最新融资之际，投资者正日益关注软银将如何为其庞大的AI雄心提供资金，以及杠杆上升对其资产负债表的影响。@@@新浪科技消息，9月23日晚间，小米在秋季新品发布会上正式发布小米18 Pro、小米18 Pro Max，并同步推出小米平板9系列、小米手环11、小米手表S5以及多款科技家电新品。小米集团合伙人、集团总裁卢伟冰表示，从系统软件到芯片硬件，从影像到大模型应用，小米18 Pro系列是小米用AI全面改造智能手机的开始，小米数字系列迎来史上最大升级。AI能力也进一步进入小米电视、净烟机、智能门锁等科技家电，推动家电产品从“被动响应”走向“主动服务”。性能方面，小米18 Pro系列首发高通新一代骁龙2nm旗舰移动平台，18 Pro和18 Pro Max分别搭载第六代骁龙8至尊版和第六代骁龙8超级至尊版，均采用2nm制程。屏幕全系搭载超级像素2.0，将硬件级防窥能力融入屏幕，并支持自由开关。在小米17 Pro系列背屏基础上，小米18 Pro系列引入AI生成能力，推出AI百变背屏，支持宠物、IP、二次元等形象定制，并可通过AI生成动态合影；首批上线超过100款应用卡，并支持通过提示词生成专属应用卡。此前，小米17 Pro系列背屏推出后，每天近400万人使用，每月超1亿人次使用，近5亿人次添加背屏壁纸。影像方面，小米18 Pro系列升级徕卡光学三摄，主摄和长焦组成徕卡双两亿像素、双大底组合；最新推出的“传奇一瞬”功能，通过全新影像处理方案提升复杂场景下的成像表现。小米18 Pro系列还搭载融入小米MiMo大模型的超级小爱2.0，能够听懂、记住、思考并执行任务，可自主使用超过260个系统工具。产品采用0.99mm极窄四等边设计，分别配备7000mAh和8500mAh电池。此次发布会还推出多款科技家电新品。小米电视S Pro RGB-Mini LED 2027搭载AI自适应广视角功能，通过毫米波雷达识别用户观看位置，并实时进行AI色彩修正；米家智能净烟机3 Max搭载AI视觉控烟技术，通过视觉模组实时记录油烟变化，并结合VLA视觉模型进行判断；小米智能门锁5 Max内外双摄搭载AI多模态大模型，可结合AI身份识别和开锁方式判断来访者身份。小米还带来米家三筒洗衣机Pro、米家意式半自动咖啡机Pro等新品，覆盖洗护、可穿戴、客厅、厨房等生活场景。价格方面，小米18 Pro起售价5999元，小米18 Pro Max起售价6999元，小米平板9系列起售价2999元，小米手表S5起售价1299元，小米手环11起售价299元；最新款小米电视、净烟机、智能门锁、洗衣机、咖啡机起售价分别为5999元、4599元、3699元、3999元、3499元。@@@据Readhub消息，DeepSeek创始人梁文锋署名的最新论文公开，首次系统披露其面向Agent训练的沙盒平台DSec的技术细节。该平台单个生产单元由约160个CPU节点、3万核和250TB内存构成，托管PB级镜像，单日可服务约300万个沙盒，峰值并发超过38万个，沙盒创建速度超过每秒5000个，单训练任务最多可一次性拉起3.2万个沙盒。DSec已支撑DeepSeek V3.2到V4.1的全部RL训练与评测沙盒负载。针对Agent强化学习需要大量保留运行状态的沙盒、传统容器方案难以适配的问题，DSec以四大类后端统一覆盖不同场景需求，将镜像拆分为独立层按需加载，实现数万沙盒快速部署，并把rollout环节从GPU训练环境拆分独立运行；同时通过AppArmor和eBPF等机制管控Agent操作风险，形成支撑大规模Agent训练的完整基础设施方案。@@@据Readhub报道，阿里云在2026云栖大会上发布AI智能体电脑Qwen Book。该产品主打自主理解用户意图、持续学习和长期任务执行，配备多类交互入口，可支持7×24小时运行重度任务，旨在改善传统AI办公中跨应用、跨Agent协同效率不足的问题。Qwen Book基于“OS as Harness”理念，以千问端云模型为核心构建Agent计算环境，目标是演变为Token驱动、持续成长的智能体载体，推动个人智能向RSI演进。目前，该产品已与多家外部伙伴及阿里系相关应用完成适配，并开放系统接口和能力接入；阿里云还将与开源项目Omarchy合作，探索面向Agent的新一代桌面操作系统。@@@据新浪科技报道，9月24日上午，智元第20000台通用具身智能机器人远征A3 Ultra在横琴长隆飞船乐园正式交付长隆集团。该机器人将与其他智元机器人一起在园区常态化上岗，服务游客互动、导览、演艺及科普研学等场景。此次第20000台机器人量产下线并交付，标志着智元具身智能产品从研发验证和单点展示，进一步迈向批量生产、规模交付和常态化运营，也意味着通用具身智能机器人从工业、商演等相对结构化场景，进入高客流、高互动和高开放度的文旅公共场景。围绕规模化交付，智元持续推进核心零部件自研、产品平台化和全流程质量管控，每台机器人均需经过整机高温老化、连续工况疲劳测试、多场景精度校验和稳定性压力测试等出厂检测，以提升产品一致性和长期运行能力。依托“作业智能、交互智能、运动智能”三智一体技术架构，A3 Ultra将在飞船乐园参与更多动态表演、角色互动和服务任务。@@@新浪科技消息，9月24日下午，在深圳“智付前海·喜迎APEC”发布会上，腾讯携手工商银行深圳市分行、深圳通公司等启动2026“畅行中国”入境支付便利服务焕新行动，迎接APEC中国年。腾讯宣布，围绕全球支付合作升级、AI与金融科技创新、深圳多场景服务体系建设三方面提升境外宾客体验。全球支付合作上，腾讯将通过TenPay Global跨境支付平台和银联国际合作，并携手网联，依托“跨境二维码统一网关”接入更多境外钱包。AI与金融科技创新上，微信支付依托“外卡内绑”“外包内用”合作网络，推出首个面向海外用户的来华支付应用TenPayGo，一次性支持国际银行卡、ApplePay、境外钱包等，在境内数千万微信支付商家消费，并延伸至城市出行、离境退税等入境友好场景。深圳多场景服务上，腾讯将携手金融机构、公共交通支付平台、重点商圈等，把支付便利送到境外宾客抵达深圳后的更多场景。@@@好啦，今天的节目到这里就结束啦，谢谢你们的陪伴，愿今天的小小分享能带来好心情。如果您喜欢我们的节目，也非常欢迎订阅和转发给你的小伙伴，这对我们非常有帮助。那我们下期再见吧！前往小宇宙评论区与主播互动

Hi~~ Good evening, everyone! It's Wednesday, September 23, 2026, and you're listening to Good Night Coffee. I'm your host, An'an! Tonight, let's talk about the latest and juiciest business and tech news—there's plenty of substance and some fun stories too. So get comfy, and let's kick off today's show!@@@According to Readhub, Apple shares rose nearly 2%, hitting an intraday high of $345.34, a record, and its total market cap once again topped $5 trillion. JPMorgan said in a research note that delivery wait times for the iPhone 18 Pro series lengthened notably after orders entered their second week, with a bigger increase than for the iPhone 17 Pro series, possibly indicating even hotter demand. JPMorgan's data showed the average shipping wait for the iPhone 18 Pro rose from 7 days in the first week to 23 days; for the iPhone 18 Pro Max, it stretched from 19 days in the first week to 30 days in the second.@@@According to TechWeb and Kuai Technology, on September 23, OpenAI launched GPT-6 Sol and GPT-6 Luna, with API pricing slashed by 50% from the promotional price of the previous generation, GPT-5.6. Sol is positioned as mid-tier and handles complex tasks such as coding, priced at $2 per million input tokens and $10 per million output tokens; Luna targets lightweight scenarios, at $0.10 input and $0.50 output. Sol cut factual accuracy errors by about half in internal tests compared with the previous generation. Both models are available to Plus, Pro, Business, Enterprise, and Edu users and can be used in ChatGPT Work and Codex; free and Go users can access Luna through the desktop app, though it is not yet available in standard ChatGPT chat. For comparison, DeepSeek V4.1 Flash's off-peak cache-miss input price is 1 yuan per million tokens and 4 yuan for output, rising to 2 yuan and 8 yuan at peak times; for regular new requests, Luna is cheaper and has no peak/off-peak pricing swings. Also on the same day, Anthropic released Claude Opus 5.5, which reaches Claude Fable 5.1 levels on most tasks, priced at $4 per million input tokens and $20 per million output tokens—20% lower than Opus 5—with cache read prices down 60% and typical workload running costs down about 40%. OpenAI's CFO said the price cuts are meant to compete with low-cost open-source models.@@@Sina Tech reports that Royal Caribbean International announced Wednesday it will acquire a 50% stake in all-inclusive resort operator Sandals Resorts International for $3 billion, valuing the latter at $6 billion. This is the company's largest deal ever and is intended to expand from cruises at sea into land-based vacations. Royal Caribbean will hold 50% of Sandals' business through a newly formed joint venture, while some members of the Stewart family retain control of the other half. The deal is expected to close in early 2027. After the news broke, Royal Caribbean shares closed down 6.1% on Wednesday. According to Investing.com, the deal values the resort business at about 10 times expected EBITDA, and Royal Caribbean has secured committed debt financing from Morgan Stanley and expects the deal to be accretive to earnings starting next year. Royal Caribbean CEO Jason Liberty called it an "important next step" in building a broader vacation platform; he will co-manage the joint venture board with Sandals Executive Chairman Adam Stewart. Sandals has 20 beachfront hotels in the Caribbean, focuses on couples' bookings, and also operates the family brand Beaches Resorts. It is the region's largest private employer. The transaction ends Sandals' multi-year sale process; after founder Gordon Stewart died in 2021, family litigation over trust governance ensued, and multiple attempts to sell over the past decade failed.@@@According to The Verge, Qualcomm unveiled the Snapdragon 8 Elite Gen 6 and, alongside it, the Snapdragon 8 Elite Extreme Gen 6. Both are positioned as flagships and built on a 2nm process, and they are identical in modem, display, security, GPS, and memory and storage support. Qualcomm no longer discloses a detailed core breakdown, but the CPU specs are the same: two prime cores up to 5GHz and six performance cores up to 4GHz. The Extreme has slightly stronger CPU, GPU, and NPU performance, with differences mainly in video, AI, and gaming: it supports 8K 60fps and 4K 240fps capture and the Advanced Professional Video (APV) codec; its NPU has a larger shared memory pool for faster, more efficient AI; its GPU adds AI-specific matrix cores and more memory, delivering a small gaming performance boost when using AI features. Motorola is the first to confirm that its newly announced Signature 27 flagship will use the Extreme chip. Qualcomm also announced that Vivo, Xiaomi, Oppo, Honor, iQoo, Nubia, OnePlus, Redmi, RedMagic, and ZTE will also adopt the two chips. They are Qualcomm's first 2nm products and, along with Apple's A20 Pro and MediaTek's Dimensity 9600 Pro, will bring 2nm process technology to smartphones.@@@cnBeta reports that six multinational financial institutions, including Bank of America and Capital One, jointly released a report titled "Building Trust in Agentic Commerce," saying that although AI-agent shopping has potential, trust issues around transparency, security, privacy and data protection, user choice, and platform interoperability remain unresolved. Agentic shopping refers to AI agents that, with authorization, autonomously browse, compare prices, choose payment methods, and place orders. The report says that as AI agents gain more autonomy, risks grow in tandem; users worry about buying the wrong thing, overspending, scams, account hijacking, or financial loss. They are willing to use AI for advice, but most do not want it to fully control the purchase process. The report also notes that AI agents may prioritize products with higher commissions or payment channels with lower transaction costs, and may even choose non-optimal options to save on compute costs. Recently, Meta patched a zero-day vulnerability in its AI assistant Muse that could have allowed attackers to take over; Amazon accused Muse agents of accessing its platform without identifying themselves, demanded they stop, and raised concerns about how account credentials are handled. The six banks did not completely reject the model, saying it could still become one of the mainstream transaction methods, and expressed willingness to work with customers, industry partners, and regulators. The participating banking alliance has prepared next-stage research to implement five core principles and establish industry standards.@@@According to Sina Tech, Dongfeng Motor said in a September 23 statement that Dongfeng Motor General Manager and Deputy Party Secretary Feng Changjun traveled to Shenzhen for talks with Huawei founder Ren Zhengfei and senior Huawei representatives. The two sides held in-depth exchanges on deepening strategic cooperation and collaborative product and technology innovation. During the talks, they reviewed the long-standing achievements of Dongfeng-Huawei cooperation and said they would continue advancing all-round, deep cooperation in automotive intelligence, enterprise digital and intelligent upgrading, and ecosystem co-building, pushing China's smart vehicle industry into a new stage of "systematic innovation." Ren Zhengfei said the foundation of carmaking lies in safety and quality, and Huawei is willing to leverage its strengths to work with Dongfeng to create smart vehicle products that stand up to market scrutiny. Feng Changjun strongly agreed, stressing that safety and quality are the foundation for smart vehicles to win market trust. He added that the two sides still have broad room to expand in ecosystem building—they can work with power grids to explore sustainable business models for vehicle-to-grid interaction, and identify deployable cooperation scenarios in in-vehicle communications technology and the integration of cars with smart transportation and smart cities. The two sides also exchanged views on business operations and development, focusing on jointly developed strategic new models such as the Yijing X9, Mengshi X700, and Voyah Dreamer 9, and discussed ways to deepen cooperation in future product definition, intelligent technology deployment, and marketing. Representatives from Dongfeng's Yijing, Mengshi, and Voyah brands took part. Dongfeng presented Huawei with a model of the Yijing X9, which is equipped with Huawei's Qiankun full-stack intelligent automotive solution.@@@According to Readhub, at the opening ceremony of the 48th WorldSkills Competition, 19 Unitree H2 humanoid robots danced onstage with 120 dancers, completing the world's first fully AI-driven autonomous cluster performance by full-size general-purpose humanoid robots. During the performance, the robots relied on onboard LiDAR for environmental perception and centimeter-level autonomous positioning, and used Unitree's self-developed cluster coordination and scheduling system to achieve millisecond-level synchronized coordination. The robots' full-body dance movements were generated by reinforcement learning algorithms and could follow the music's rhythm. The system also has autonomous error-correction capabilities: when external interference causes a positional deviation, the robots can autonomously detect the deviation and automatically correct themselves back into place.@@@According to Readhub, Chinese AI startup DeepSeek will brief the UN Security Council this week on the risks posed by artificial intelligence. The UN Security Council is scheduled to meet locally on September 23 to discuss AI and international security. Weeks earlier, several AI industry leaders called for coordinated efforts to slow the development of increasingly powerful AI systems, warning that such systems could soon improve themselves and slip beyond human control. Several Chinese AI companies, including DeepSeek and Moonshot AI, were invited to make statements at the meeting. However, Deep

嗨~~大家晚上好呀！今天是2026年9月23日，星期三，欢迎来到晚安咖啡，我是主播安安！今晚咱们一起聊聊新鲜热乎的商业科技动态，有干货也有趣闻，快找个舒服的姿势，跟我开启今天的分享吧！@@@据Readhub报道，苹果股价上涨近2%，盘中最高触及345.34美元，创历史新高，总市值再次突破5万亿美元大关。摩根大通发布研报称，iPhone 18 Pro系列订单进入第二周后，交付等待周期明显拉长，且拉长幅度超过iPhone 17 Pro系列，或表明市场需求更火热。研报统计显示，iPhone 18 Pro第二周平均发货等待时长从首周的7天增至23天；iPhone 18 Pro Max则从首周的19天拉长至次周的30天。@@@据TechWeb、快科技报道，9月23日，OpenAI发布GPT-6 Sol和GPT-6 Luna，API定价较上一代GPT-5.6促销价直降50%。Sol定位中阶，承担编程等复杂任务，每百万token输入2美元、输出10美元；Luna面向轻量场景，输入0.10美元、输出0.50美元。Sol内部事实准确性测试错误较上一代减少约一半。两款模型已向Plus、Pro、Business、Enterprise及Edu用户开放，可在ChatGPT Work和Codex使用；免费版和Go版用户可通过桌面端使用Luna，尚未在ChatGPT普通聊天开放。作为对照，DeepSeek V4.1 Flash闲时缓存未命中输入价每百万token 1元、输出4元，高峰升至2元和8元；按普通新请求计算，Luna更便宜且无峰谷定价波动。同天，Anthropic发布Claude Opus 5.5，多数任务达到Claude Fable 5.1水平，每百万输入token 4美元、输出20美元，较Opus 5下降20%，缓存读取价格降60%，典型工作负载运行成本降低约40%。OpenAI首席财务官表示，降价是为与低成本开源模型竞争。@@@新浪科技消息，皇家加勒比国际游轮公司周三宣布，将以30亿美元收购全包式度假村运营商Sandals Resorts International 50%股权，对后者估值60亿美元。这是公司史上最大交易，意在从海上扩展至陆地度假市场。皇家加勒比将通过新设合资企业持有Sandals业务50%股权，斯图尔特家族部分成员保留另一半控制权，交易预计2027年初完成。消息公布后，其股价周三收跌6.1%。据Investing.com报道，该交易对度假村业务估值约为预期EBITDA的10倍，皇家加勒比已获摩根士丹利承诺性债务融资，预计明年起增厚盈利。皇家加勒比CEO贾森·利伯蒂称这是构建更广泛度假平台的“重要下一步”，他将与Sandals执行主席亚当·斯图尔特共同管理合资公司董事会。Sandals在加勒比地区拥有20家海滨酒店，主打情侣双人预订，并运营家庭品牌Beaches Resorts，是当地最大私营雇主。此次交易结束了Sandals多年出售进程；创始人戈登·斯图尔特2021年去世后，家族因信托治理引发诉讼，过去十年多次出售未果。@@@据The Verge报道，高通发布骁龙8 Elite Gen 6，并同步推出骁龙8 Elite Extreme Gen 6。两款均定位旗舰并采用2纳米制程，且在调制解调器、显示、安全、GPS及内存和存储支持上相同。高通不再公布核心详细拆解，但CPU规格一致：两大核最高5GHz，六性能核最高4GHz。Extreme的CPU、GPU、NPU性能略强，差异主要在视频、AI和游戏：支持8K 60fps、4K 240fps拍摄及高级专业视频（APV）编解码器；NPU共享内存池更大，AI更快更高效；GPU加入AI专用矩阵核心和更多内存，使用AI功能时小幅提升游戏性能。摩托罗拉首个确认，新宣布的Signature 27旗舰将采用Extreme芯片。高通还宣布，Vivo、小米、Oppo、荣耀、iQoo、努比亚、一加、Redmi、红魔和中兴也将采用这两款芯片。两款芯片是高通首批2纳米产品，与苹果A20 Pro、联发科天玑9600 Pro一同将2纳米制程带入智能手机。@@@cnBeta消息，美国银行、第一资本银行等六家跨国金融机构联合发布《建立代理式商业信任》报告，称AI代理购物虽具潜力，但透明度、安全、隐私与数据保护、用户选择权和平台互操作等信任危机待解。代理式购物指AI代理经授权自主浏览、比价、选择支付并下单。报告称，随着AI代理自主权扩大，风险同步放大；用户担心买错、超预算、诈骗、账户劫持或资金损失，愿用AI获取建议，但多数不愿其完全掌控购买流程。报告还指出，AI代理可能优先推荐佣金更高商品、交易成本更低支付渠道，甚至为省计算费选择非最佳方案。近期，Meta修复AI助手Muse中可让攻击者接管的零日漏洞；亚马逊指责Muse代理访问其平台未表明身份，要求停止并担忧账户凭证处理。六家银行未完全否定该模式，认为其仍可能成为主流交易方式之一，并愿与客户、产业伙伴及监管机构合作。参与银行联盟已准备下一阶段研究，以落实五项核心原则并建立行业标准。@@@据新浪科技消息，东风汽车9月23日发文称，东风汽车总经理、党委副书记冯长军赴深圳，与华为创始人任正非及华为高层代表会谈。双方围绕深化战略合作、产品技术协同创新等议题深入交流。会谈中，双方回顾了东风与华为长期积淀的合作成果，并表示围绕汽车智能化、企业数字化和智能化升级、生态共建等方向持续推进全方位深度合作，推动中国智能汽车产业步入“体系化创新”新阶段。任正非表示，造车之本在于安全与品质，华为愿发挥所长，与东风共同打造经得起市场检验的智能汽车产品。冯长军对此表示高度认同，强调安全与品质是智能汽车赢得市场信任的基础，双方在生态建设领域仍有广阔拓展空间，可携手电网探索车网互动可持续商业模式，并在车载通信技术、汽车与智慧交通、智慧城市融合等方向挖掘合作落地场景。双方还就企业经营、业务发展情况沟通交流，重点围绕奕境X9、猛士X700、岚图梦想家9等联合打造的战略新车型交换看法，并聚焦未来产品定义、智能技术落地、市场营销等维度探讨深化合作方向。东风汽车旗下奕境、猛士、岚图品牌代表参与本次交流，东风现场向华为赠送奕境X9车模，该车搭载华为乾崑全栈智能汽车解决方案。@@@据Readhub消息，在第48届世界技能大赛开幕式上，宇树科技19台H2人形机器人与120名舞者同台共舞，完成全球首次全尺寸通用人形机器人全AI驱动自主集群表演。表演中，机器人依靠机载激光雷达实现环境感知与厘米级自主定位，并借助宇树自研的集群协同调度系统实现毫秒级协同同步。机器人的舞蹈全身动作由强化学习算法生成，能够跟随音乐节奏。整套系统还具备自主纠错能力，当受到外界干扰出现位置偏移时，机器人可自主识别偏差并自动修正归位。@@@据Readhub消息，中国人工智能初创企业DeepSeek将在本周向联合国安理会通报人工智能带来的风险。联合国安理会定于当地时间9月23日召开会议，讨论人工智能与国际安全问题。此前数周，多位人工智能行业领袖呼吁协调放缓日益强大的人工智能系统的发展步伐，并警告这类系统可能很快实现自我改进，进而脱离人类控制。包括DeepSeek和月之暗面在内的多家中国人工智能企业受邀在会上发表声明。不过，DeepSeek创始人梁文锋没有计划出席此次会议，相关安排仍可能发生变化。@@@量子位消息，林俊旸离职半年后，阿里千问新任负责人确定，由前华为“天才少年”刘大一恒接棒。阿里云栖大会官网显示，他任阿里巴巴ATH事业群Token Foundry Qwen LLM项目负责人，排序紧随蔡崇信、吴泳铭；他在吴泳铭演讲后作15分钟主题演讲《Qwen：迈向真实世界智能体》。刘大一恒是四川大学计算机学院本硕博，2021年加入达摩院，负责Qwen最早预训练，累计参与300多款Qwen模型开源；谷歌学术引用超六万次，以通讯作者获NeurIPS 2025最佳论文奖，另一通讯作者是林俊旸。2020年他成为西南首位入选华为天才少年第一档学生，年薪200万元，但未入职，后以阿里星加入阿里。今年3月林俊旸宣布卸任，阿里云CTO周靖人代管Qwen一号位；刘大一恒在预训练外兼管后训练和Coding，向周汇报。6月通义大模型事业部与未来生活实验室合并为Token Foundry，由吴泳铭直管，周靖人转任首席科学家。刘近期全面接管Qwen。Qwen3.8 Max在Artificial Analysis智能指数排第七，仅次于Anthropic、OpenAI旗舰模型及Meta的Muse Spark 1.3。@@@新浪科技消息，飞行汽车企业汇天与阿联酋拉斯海玛酋长国正式签署战略合作协议，将共建全球首个个人低空飞行监管沙盒，即阿联酋先进空中交通监管沙盒平台，并由阿联酋民航局全程监管，同步启动阿布扎比—拉斯海玛跨酋长国航线规划与示范落地。该平台由阿联酋民航总局统筹监管，是阿联酋官方认证的专业化空中交通监管试验平台，旨在探索可复制推广的飞行汽车适航监管、安全运行和商业运营体系。汇天创始人赵德力称，公司深耕中东五年，已验证中国飞行汽车具备出海技术实力，能适配国际航空监管规则，未来将推进全球化战略。依托沙盒计划，上述跨酋长国示范航线将开展常态化试飞与试运行，这是中东跨行政区空中交通网络化运营的探索。近日，汇天受邀参加中国驻阿联酋大使馆庆祝中华人民共和国成立77周年招待会；在驻阿大使曾继新和阿联酋王室成员见证下，“陆地航母”在阿布扎比酋长国宫完成双机编队飞行。@@@好啦，今天的节目到这里就结束啦！谢谢大家的陪伴，愿今晚的梦甜甜的。如果您喜欢我们的节目，也非常欢迎订阅和转发给你的小伙伴，这对我们非常有帮助。那我们下期再见吧！前往小宇宙评论区与主播互动

嗨~~大家晚上好呀！今天是2026年9月22日，星期二，我是晚安咖啡的主播安安。夜色刚好，故事正热，今天咱们一起聊聊最新鲜的商业科技动态，看看有哪些新变化、新机会。准备好，跟我一起开启今晚的精彩吧！@@@据Readhub报道，Meta推出的个人AI Agent产品Muse上线约10天，登顶美国App Store免费应用榜。Muse通过独立虚拟机、Sentinel安全Agent、支付隔离等多重机制保障安全，并依托专为个人Agent打造的Muse Spark模型，以及Meta自身的社交产品生态实现快速走红。此前，Meta在大模型竞赛中进展有限，收购Manus受阻后转向个人Agent赛道，不再执着于SOTA模型，转而以模型为底座发力应用层。当前科技巨头在AI领域路径分化：谷歌坚持全栈打法；微软、亚马逊聚焦企业服务与AI基础设施，打造多模型平台；Meta则依托消费级用户和社交生态优势深耕个人Agent。国内腾讯、字节、阿里也基于各自优势布局个人Agent，分别从关系服务、内容推荐、交易商业服务等切入。此外，个人Agent赛道的Manus、Instinct也获资本青睐。@@@据虎嗅24小时消息，Anthropic首席执行官达里奥·阿莫迪计划于周三与OpenAI首席执行官萨姆·奥尔特曼一同向联合国安理会进行简报，此次简报的主题聚焦人工智能的未来发展。@@@cnBeta消息，黄仁勋已确认出席中美领导人国宴。该报道同时称，英伟达GPU重返中国市场有戏。@@@据Readhub消息，谷歌推出全新高端笔记本电脑产品线Googlebook，起售价899美元。该产品线主要面向安卓手机用户和程序员，谷歌希望让Gemini人工智能助手成为其使用体验的核心，并借此打造类似苹果的设备协同生态。谷歌同时明确，Googlebook并非用于取代现有的Chromebook，也不是一款面向所有用户的通用笔记本。@@@据CNBC报道，特斯拉本月在越南注册“特斯拉汽车越南公司”，商业登记记录显示，此举为其潜在进入越南市场铺路。越南2025年成为东南亚最大电动车市场，销量翻倍以上，电动车占新车销售近40%，据国际能源署5月数据。HSC研究显示，本土车企VinFast已占越南国内电动车市场92%。VinFast由Vingroup支持，后者2026年上半年营收221.97万亿越南盾，约85.2亿美元。BMI旗下Fitch Solutions分析师称，VinFast的本地品牌认知、Vingroup消费生态、关联电动出租车网络及充电售后体系，降低了消费者持有风险。Mobility Global东盟分析师称，VinFast有超过15万个仅限其车辆的专有充电端口，也给比亚迪在越南带来挑战。特斯拉需建立分销和服务覆盖，并建立充电信心；当地买家对价格和实用性高度敏感。潜在入市时间尚未公开。报道指出，与泰国、印尼不同，越南已有本土冠军和快速扩张的出行生态。@@@据TechWeb报道，9月22日，在2026杭州云栖大会上，平头哥发布训推一体AI芯片真武V900。其单芯片性能为真武M890的3倍，配216GB显存，片间互联带宽1200GB/s，原生支持FP8、FP4，面向万亿级模型训练与推理。V900计划2027年一季度量产并在阿里云数据中心规模化部署；2027年三季度发布真武J900。它通过自研ICN Switch实现千卡全带宽互联，磐久超节点服务器集成V900、ICN Switch、磐脉智能网卡及镇岳SSD主控，结合阿里云新网络架构，单一AI集群可扩至50万卡。基于M890的超节点已大规模应用，跑通Qwen3.8、Kimi K3等超2万亿参数模型；今年四季度阿里云将新增服务节点。截至2026年6月，真武系列已服务超650家企业客户。平头哥还计划2027年推出倚天720和730，730首次采用全自研微架构，单核SPECint 2017/GHz性能最高达倚天710的1.4倍；2029年推出倚天750，可与真武AI芯片经ICN总线直连。@@@据Readhub消息，知情人士透露，DeepSeek周日召开了一场闭门线下会议，要求投资者前往北京或杭州办公室参会，CEO梁文锋以线上方式接入。为防止信息泄露，参会投资者被要求上交电子设备及随身包袋，现场仅提供纸笔记录。梁文锋在会上称，公司正在训练一个2万亿参数模型，规模超过现有1.4万亿参数的V4旗舰模型；此外，DeepSeek还计划开发8万亿参数模型。知情人士称，该公司正在加入大模型开发竞赛。@@@TechWeb援引TomsHardware报道，华为已暂时搁置全球AI芯片推广计划。华为副董事长、轮值董事长汪涛在上海举行的2025年华为全联接大会上透露，由于国内算力需求已超过供给，没有足够产能满足国内需求，公司暂时还没有全面进军国际市场的计划，但会向部分需求特别旺盛的国家供应少量产品。今年初在西班牙巴塞罗那举行的MWC2026期间，华为首次在海外展示最新超节点产品和解决方案，涵盖多个型号，并表示通过“集群+超节点”架构创新打造算力底座，坚持开源开放。此后有消息称，华为计划2026年第四季度进入韩国AI芯片市场，这将是其首次大规模进军海外市场。根据最新披露，华为下一代昇腾900系列AI加速器只面向中国销售，不进入国际市场，公司本周确认了这一安排；即将推出的昇腾960系列NPU在能力上可媲美AMD与英伟达现有部分AI GPU，但海外需求本身也存在不确定性。@@@据Readhub消息，宇树科技发布了Dex5-S灵巧手。@@@据新浪科技消息，在9月22日上午举行的2026美团即时零售产业大会上，美团核心本地商业CEO王莆中发表演讲。他回顾即时零售行业过去五年发展时透露，美团2021年提出的“即时零售市场规模5年达到1万亿、交易用户规模超过5亿”的展望目标已提前完成。王莆中表示，即时零售不是应急零售，而是一种高确定性的生活方式，从鲜花、酒水到3C数码，从定制产品到新型商超推出的产品，这种生活方式已走入千家万户。他还提到，美团2024年提出到2027年推动建设10万家闪电仓、达到2000亿GMV，从今天掌握的数据来看，这一目标一定可以提前完成。谈及行业增长飞轮，他将其简化为用户需求、基建建设和商品供给三个关键要素，认为需求越高频、容量越大，流通网络和基建越成熟，商品创新空间就越大。@@@好啦，今天的节目到这里就结束啦。谢谢你们的陪伴，愿好心情一直陪着你。如果您喜欢我们的节目，也非常欢迎订阅和转发给你的小伙伴，这对我们非常有帮助。那我们下期再见吧！前往小宇宙评论区与主播互动