广告
加载中

库克的葫芦里 没有苹果的药

赵卫卫 2026-06-10 13:16
赵卫卫 2026/06/10 13:16

邦小白快读

EN
全文速览

本文核心披露了2026苹果开发者大会上苹果AI转型的最新进展,普通读者可获取以下核心干货:

1. 新一代Siri AI推进情况:推迟两年后大概率于今年9月推出beta版本,首发地区不包含中国和欧盟,该产品由苹果与谷歌联合研发,苹果仅获得Gemini模型授权,在此基础上自研部分功能。

2. 新Siri AI的功能亮点:支持屏幕感知、多设备同步,手机端iOS27从灵动岛下滑即可唤起,相机新增Siri识别模式;Mac端整合进Spotlight搜索框;Safari新增AI标签整理、网页变化提醒等功能,同时实现全系统跨应用AI联动,还优化了未成年用户的家长管控功能。

3. 对现有苹果用户的影响:iPhone16系列用户无法适配苹果性能最强的本地AI模型,仅12G内存的新款机型可支持,苹果此前的AI宣传透支用户信任,且联合谷歌研发的模式也让苹果隐私保护人设受到质疑。

本文为消费电子品牌的AI转型提供了多维度参考干货,具体内容如下:

1. 用户认知与信任层面:当前消费者既关注AI功能落地,也重视隐私保护,对品牌兑现承诺的要求极高,苹果多次推迟AI落地、旧旗舰无法享受顶配AI体验的操作,已经引发大规模用户信任危机,2024年还因AI功能虚假宣传赔付2.5亿美元,提醒品牌切忌透支信任做超前营销。

2. AI转型策略选择:苹果选择租赁谷歌大模型控制转型成本,每年仅需投入10亿美元,远低于自研大模型每年上百亿的投入,这种轻资产合作转型的策略,给资金有限的品牌提供了新思路,但也要注意,和对手合作研发会引发隐私层面的用户质疑,需要提前做好公众沟通。

3. 产品研发方向:苹果将AI能力嵌入系统全场景各个环节,同时兼顾普通用户和未成年群体的细分需求,这种全链路适配+覆盖细分人群的研发思路值得品牌参考。

对于3C数码产品卖家,本文披露的苹果AI转型动向,带来了以下机会提示与风险预警干货:

1. 市场需求与机会:新一代Siri AI beta版将于今年9月推出,仅12G内存的新款iPhone可支持顶配AI功能,国内消费者对苹果AI功能有较高期待,将会带动新款iPhone的换机需求,卖家可提前布局12G内存新款机型的备货,抢占市场先机。

2. 风险提示:iPhone16系列此前宣传主打AI能力,但如今无法适配顶配本地AI模型,已经引发大量老用户不满,卖家需要提前做好用户咨询应对方案,梳理相关问题的回复口径,避免引发售后纠纷和负面口碑。

3. 后续动向提示:苹果即将换帅,新任CEO将在9月iPhone18发布会上正式登场,后续苹果AI转型策略可能发生调整,卖家需要持续跟进官方动向,及时调整自身的销售和备货策略。

对于苹果供应链工厂及消费电子代工厂,本文披露的信息带来以下干货参考:

1. 产品生产需求变化:苹果新一代支持顶配AI功能的机型,将内存门槛提升到12G,未来苹果旗舰机型的基础内存配置会全面升级,上游工厂需要提前调整生产方案,适配新的硬件参数要求,提前做好产能规划。

2. 潜在商业机会:苹果AI转型起步晚,基础设施投入不足,当前高度依赖谷歌的模型和云服务,后续苹果一定会加大本地AI相关芯片、硬件的研发生产投入,给上游合作工厂带来更多订单和合作机会,工厂可提前做好技术和产能储备。

3. 数字化转型启示:苹果将AI能力全面嵌入产品交互、功能设计等各个环节,提示上游制造工厂可以加快自身数字化转型,借助AI技术优化生产流程、提升产品设计能力,增强自身的市场竞争力。

对于AI服务商、云服务商等科技服务商,本文披露的行业动向带来以下干货参考:

1. 行业发展趋势:当前前沿大模型自研成本极高,头部企业自研年投入可达150亿美元,越来越多科技巨头倾向选择租赁合作模式,而非投入巨资自研,这给大模型服务商、云服务商带来了大量B端合作的市场机会。

2. 客户核心痛点:科技企业AI转型的核心痛点一是成本控制,二是平衡隐私保护诉求,苹果和谷歌的合作模式,既控制了年10亿美元的转型成本,又通过数据脱敏保留了自身的隐私品牌人设,这种模式可以作为服务商给客户设计解决方案的参考。

3. 产品研发方向:企业客户同时需要云端高性能大模型和终端适配的本地轻量模型,苹果同时布局本地端、云端共五款分层模型对应不同场景,说明端云结合的AI方案是未来发展方向,服务商可针对性开发相关产品。

对于开发者平台等科技平台商,本文披露的苹果WWDC大会动向带来以下干货参考:

1. 平台用户核心需求:当前平台用户对原生AI功能的需求非常强烈,期待平台推出成熟落地的AI能力,同时用户对隐私保护、平台兑现产品承诺的要求越来越高,平台运营需要重视这类核心诉求。

2. 平台AI生态布局动向:苹果本次开发者大会开放了多项AI能力,支持自然语言创建快捷指令、Safari生成浏览器扩展,给开发者留出了更多AI开发空间,可见头部科技平台会持续加大AI生态的布局,逐步开放AI基础能力给生态伙伴。

3. 风险规避提示:苹果此前提前宣传未落地的AI功能,引发集体诉讼最终赔付2.5亿美元,又因为旧机型不支持新功能引发用户信任危机,提示平台不要过度超前宣传未落地的功能,需要平衡技术迭代速度和对用户的承诺,避免引发信任危机和法律风险。

对于科技产业研究者,本文披露了苹果AI转型的最新动向,带来不少值得研究的干货内容:

1. 产业新动向:全球头部消费电子企业在AI浪潮下转型路径出现分化,谷歌、微软等巨头已经在AI领域构建起技术护城河,苹果作为全球盈利能力最强的消费电子巨头,转型步伐滞后,选择了成本优先的合作模式,租用对手大模型而非自研,这是头部企业AI转型的一种全新路径,值得深入研究。

2. 产业新问题:这种跨对手合作的模式带来了新的隐私和信任问题,苹果虽然对数据做了脱敏处理,但核心运算依托谷歌基础设施,隐私保护的边界存在模糊地带;同时手机换机周期和AI技术迭代速度不匹配,品牌如何兑现AI承诺成为整个行业需要解决的新问题。

3. 商业模式层面:苹果采用端云结合的分层模型布局,用不同性能的模型适配不同价位硬件,既拉动了新硬件的销售,又控制了AI转型的整体成本,这种商业模式的创新值得研究者关注探讨。

返回默认

声明:快读内容全程由AI生成,请注意甄别信息。如您发现问题,请发送邮件至 run@ebrun.com 。

我是 品牌商 卖家 工厂 服务商 平台商 研究者 帮我再读一遍。

Quick Summary

This article reveals the latest updates on Apple's AI transformation showcased at the 2026 Apple Worldwide Developers Conference. General readers can get the following key takeaways:

1. Development progress of the new-generation AI-powered Siri: After a two-year delay, a beta version is highly likely to launch this September, with China and the European Union excluded from the first launch regions. The product is co-developed by Apple and Google; Apple only secured licensing for Google's Gemini model, and developed some of its own features on top of this foundation.

2. Key feature highlights of the new AI Siri: It supports on-screen context awareness and multi-device syncing. On iOS 27 for iPhones, it can be invoked by swiping down from the Dynamic Island, and the Camera app gets a new Siri recognition mode. On Macs, it is integrated into the Spotlight search bar. Safari adds new AI-powered features including tab organization and webpage change alerts, along with system-wide cross-app AI integration. Parental control features for underage users have also been optimized.

3. Impact on existing Apple users: The iPhone 16 series will not be compatible with Apple's most powerful on-device AI model; only new models equipped with 12GB of RAM can support the full feature set. Apple's earlier hyped AI marketing has eroded user trust, and its co-development partnership with Google has also sparked questions over Apple's long-standing brand positioning around privacy protection.

This article provides multi-dimensional insights for consumer electronics brands pursuing AI transformation, as outlined below:

1. User perception and trust: Today's consumers care deeply about both practical AI feature delivery and privacy protection, and hold brands to high standards for delivering on promised features. Apple's repeated AI launch delays and the exclusion of older flagship models from top-tier AI experiences have already triggered a large-scale user trust crisis. In 2024, Apple paid $250 million to settle a lawsuit over false advertising of AI features, a reminder to brands that overhyping unlaunched products to透支 user trust is a risky strategy.

2. AI transformation strategy: Apple opted to lease a large language model from Google to control transformation costs, with an annual investment of only $1 billion, far lower than the tens of billions of dollars required for in-house large model development. This asset-light partnership strategy offers a new approach for brands with limited capital. However, brands should note that co-developing with a competitor will trigger user concerns over privacy, and require proactive public communication in advance.

3. Product R&D direction: Apple embeds AI capabilities across every system-level scenario, while catering to the segmented needs of both general users and underage users. This full-stack integration plus segmented user coverage approach is a valuable reference for other brands.

For 3C electronics sellers, the Apple AI transformation trends revealed in this article offer the following opportunity insights and risk warnings:

1. Market demand and opportunities: The beta version of the new AI-powered Siri will launch this September, and only new iPhones with 12GB of RAM can support the full suite of top-tier AI features. Chinese consumers have high expectations for Apple's AI features, which will drive upgrade demand for new iPhones. Sellers can prepare inventory of 12GB RAM new models in advance to seize first-mover advantage.

2. Risk warning: The iPhone 16 series was previously marketed as AI-focused, but it will not support the full top-tier on-device AI model, which has already sparked discontent among early adopters. Sellers should prepare response guidelines for user inquiries in advance and standardize reply scripts to avoid after-sales disputes and negative word-of-mouth.

3. Future trend reminder: Apple is preparing for a leadership change, and the new CEO will officially take the stage at the September launch event for the iPhone 18. Apple's AI transformation strategy may see adjustments going forward, so sellers should continue tracking official updates and adjust their sales and inventory strategies accordingly.

For Apple supply chain factories and consumer electronics contract manufacturers, the information revealed in this article offers the following key insights:

1. Changes in production demand: Apple has raised the RAM threshold to 12GB for its new generation of models that support full top-tier AI features, and the base RAM configuration for future Apple flagship models will see a full upgrade. Upstream factories need to adjust production plans in advance to adapt to new hardware requirements, and make capacity arrangements ahead of time.

2. Potential business opportunities: Apple started its AI transformation late, and lacks sufficient in-house infrastructure, so it currently relies heavily on Google's models and cloud services. Apple will definitely increase R&D and production investment for on-device AI-related chips and hardware going forward, which will bring more orders and partnership opportunities to upstream cooperating factories. Factories can prepare technical and capacity reserves in advance.

3. Digital transformation inspiration: Apple has fully embedded AI capabilities into all links including product interaction and feature design, which serves as a reminder for upstream manufacturing factories to accelerate their own digital transformation, leverage AI to optimize production processes and improve product design capabilities, and boost their overall market competitiveness.

For technology service providers including AI and cloud service providers, the industry trends revealed in this article offer the following key insights:

1. Industry development trends: The current R&D cost for cutting-edge in-house large models is extremely high, with annual investment reaching $15 billion for leading tech companies. A growing number of tech giants prefer the leasing partnership model over investing heavily in in-house development, which creates substantial B2B cooperation opportunities for large model and cloud service providers.

2. Core pain points of enterprise clients: The two core pain points for tech enterprises' AI transformation are cost control and balancing privacy protection demands. The Apple-Google partnership keeps annual transformation costs at $1 billion, while preserving Apple's privacy-focused brand positioning through data desensitization. This model can serve as a reference for service providers when designing customized solutions for clients.

3. Product R&D direction: Enterprise clients need both high-performance cloud large models and lightweight on-device models adapted for end devices. Apple has rolled out a layered architecture of five models across on-device and cloud settings to fit different use cases, which indicates that a hybrid on-device-cloud AI solution is the future development direction. Service providers can develop targeted products aligned with this trend.

For technology platform operators including developer platforms, the WWDC trends revealed in this article offer the following key insights:

1. Core demands of platform users: Platform users currently have very strong demand for native AI functionality, and expect platforms to launch mature, fully delivered AI capabilities. At the same time, users are increasingly demanding around privacy protection and platforms delivering on product promises, so platform operations must prioritize these core demands.

2. AI ecosystem layout trends: Apple opened up multiple AI capabilities at this year's developer conference, supporting natural language-based Shortcuts creation and Safari browser extension generation, leaving more room for AI development for third-party developers. This shows that leading technology platforms will continue to expand AI ecosystem layout, and gradually open up foundational AI capabilities to ecosystem partners.

3. Risk mitigation reminder: Apple's earlier overpromotion of unlaunched AI features led to a class-action lawsuit and a $250 million settlement, and the exclusion of older models from new AI features triggered a user trust crisis. This reminds platforms to avoid overhyping unlaunched features prematurely, and balance the pace of technological iteration with commitments to users to avoid trust crises and legal risks.

For technology industry researchers, this article reveals the latest developments in Apple's AI transformation, offering many valuable research insights:

1. New industry trends: Global leading consumer electronics companies are following divergent transformation paths amid the AI wave. While giants like Google and Microsoft have already built technical moats in the AI sector, Apple — the most profitable consumer electronics giant globally — has lagged in its transformation, and chose a cost-first partnership model by leasing a large model from a competitor rather than building it in-house. This represents a completely new path for AI transformation for leading enterprises, and deserves in-depth research.

2. New industry issues: This cross-competitor partnership model creates new privacy and trust challenges. Although Apple has desensitized user data, core computing still relies on Google's infrastructure, leaving the boundary of privacy protection ambiguous. Additionally, the length of smartphone upgrade cycles does not match the rapid iteration pace of AI technology, so how brands can deliver on AI promises has become a new industry-wide problem to solve.

3. Business model innovation: Apple has adopted a hybrid on-device-cloud layered model architecture, using models of different performance levels to match hardware at different price points. This approach both drives sales of new hardware and controls overall AI transformation costs, and this business model innovation deserves further attention and exploration from researchers.

Disclaimer: The "Quick Summary" content is entirely generated by AI. Please exercise discretion when interpreting the information. For issues or corrections, please email run@ebrun.com .

I am a Brand Seller Factory Service Provider Marketplace Seller Researcher Read it again.

苹果的AI转型,什么时候才能跟上世界的节奏?

撰文|赵卫卫

早期热度一度推动股价上涨,但正式版没有明确的上线日期,推迟了两年的苹果智能,大概率会在今年9 月推出「beta」版本,首发地区不会有中国和欧盟,这跟用户的心理预期,还有相当大的落差。

2026苹果开发者大会落幕,苹果股价大跌近2%,创下近三周以来最大单日跌幅,这也是上述问题引发的主要结果。

「只要隐私得到优先保障,我并不介意 Siri开发速度慢一些。如果我想要更快的体验,我会转投安卓阵营,把所有个人信息都交给谷歌。」《华尔街日报》的相关文章下,一则热门评论说。

但很快有用户提出反驳:苹果已与谷歌达成合作协议,下一代Siri将基于谷歌Gemini模型及云技术搭建,高阶Siri请求会转发至谷歌云服务器,这也意味着苹果的隐私保护体系无法绕开谷歌。

其核心论据为:「谷歌每年向苹果支付200亿美元,只为换取默认搜索引擎权限。这足以体现用户数据对谷歌的价值 —— 倘若苹果真能彻底守住隐私,谷歌又为何愿意持续支付这笔费用?」

这类观点道出了大批苹果用户的失望,大家质疑的并非单纯的技术问题,而是双方背后的信任危机。

尽管苹果在发布会上明确表态:「隐私在AI领域不可谈判(Privacy in AI is nonnegotiable)」,但新一代Siri由苹果联合谷歌共同研发,苹果仅获得Gemini模型授权,并在此基础上自研模型。

2026苹果开发者大会,也成为库克作为CEO的谢幕演出。他将苹果打造为全球盈利能力最强的消费电子巨头,服务业务更是构筑起难以被撼动的竞争护城河。

但这个帝国,在今天的AI浪潮中反应迟缓,需要借助对手的模型构建自己的人工智能系统。

接替库克出任苹果CEO的,是打造Max电脑和iPhone研发出身的约翰·特纳斯,面对谷歌、微软、Meta等对手在软件和人工智能领域上不断构建的护城河,特纳斯需要回答库克没有写完的答卷:苹果的AI转型,什么时候才能跟上全球行业的节奏?

为迟到的Siri做铺垫

Siri AI是苹果全新升级的语音助手,底层依托苹果与谷歌联合打造的新一代Apple Intelligence平台运行。

它有全新的外观和声音,屏幕顶部会出现一个Siri气泡,用户可以调整Siri的表达方式和语速。Siri具备屏幕感知能力,可读取当前界面内容、结合上下文作答,还内置图像编辑与写作辅助工具。相关功能支持多设备同步,交互记录会保存在全新应用Siri AI当中。

手机端搭载最新iOS 27系统,从灵动岛下滑即可唤起Siri;相机应用新增「Siri模式」,能够实时识别取景画面中的物体并展示相关信息。

Mac电脑端的最大变化,是将Siri直接整合进Spotlight搜索框。

用户可以在搜索栏里直接提问,也可以选择特定文件进行查询,无需再单独呼出助手界面。与 iOS 27一样,macOS 27同样引入了液态玻璃透明度滑块,并收窄了窗口圆角半径,同时恢复了彩色边栏图标。

值得一提的是苹果自带浏览器的升级,也让AI能力进一步落地。

比如,新增的 AI标签页管理功能可以自动按主题对所有打开的网页进行归类整理。「通知我」功能允许用户让Safari监视某个网页的变化,例如关注的商品重新上架时主动提醒。此外,用户还可以用自然语言描述需求,让Safari直接生成浏览器扩展插件。

对苹果而言,AI更深层的价值体现在跨应用联动能力上。

例如信息应用会根据对话内容,主动推荐可分享的照片;通话界面会自动展示航班确认码等相关信息;邮件功能新增智能内容建议,还支持用自然语言创建日程并同步至日历。系统快捷指令也开放了自然语言创建功能,图像游乐场则新增了构图调整和更强的「干扰物去除」能力,支持对照片进行 AI级别的二次编辑。

当然,苹果依旧兼顾全体用户,并非只面向AI爱好者。针对未成年用户,苹果也优化了家长管控功能,这也是不少家长关注的重点。

苹果对儿童账户体系进行了系统性重设计,家长可以为孩子精确选择允许使用的应用和可访问的内容,并要求孩子在打开新网站前必须申请授权。屏幕时间菜单界面也经过重新设计,操作路径更加清晰,旨在让家长能更高效地管理孩子的设备使用时长。

不难看出,苹果系统层面的改造,都在为即将到来的Siri AI做铺垫。

iPhone 16用户被背刺

这是库克最后一次以 CEO身份站在苹果WWDC的舞台上。

他并未用数据总结自己的任职生涯,而是这样说道:打造全球顶尖产品、丰富大众生活体验,始终是我们的初心。能和这群富有创造力的伙伴同行,让科技融入日常生活,是我毕生的荣幸。

与之成为反差的是,这场发布会被外界批评为「最令人失望的一次WWDC」。

「我从 2010年代初就开始关注苹果的发布会,而这绝对是迄今为止最差的一次。」「所以他们真的又在做一个非直播的演示?这不就是2024年把他们搞砸的那个东西吗?」苹果忠实用户的论坛macrumors里,有人重提了两年前的往事。

在2024年 WWDC大会上,现场展示的Siri功能并非真机实时运行,而是提前录制好的视频。彼时Siri团队内部都未曾见过可正常运行的正式版本。此后苹果因虚假宣传遭到集体诉讼,并在2025年以2.5亿美元达成庭外和解。

两年过去了,用户对苹果新的Siri吐槽的点在哪里?

Reddit论坛的留言,道出了不少苹果用户的心声。一条数百点赞的评论写道:「实在有些失望,我的iPhone 16 Pro Max无法搭载性能最强的本地Foundation Model。」

Phone 16 Pro Max是苹果2024年 9月推出的旗舰机型,当时宣传主打人工智能能力,并称iPhone 16是首款为Apple Intelligence量身打造的机型。

时隔一年半,苹果AI相关功能迟迟未能全面落地。这款机型虽可联网使用Siri AI,却无法适配苹果性能最强的本地模型。

本地模型即AI模型直接在手机芯片上运行,无需联网、数据全程留存设备内,这也是苹果主打的隐私优势。运行苹果最强力的本地模型,需要的硬件门槛是12G的内存,满足这条的手机型号,目前只有iPhone Air(12GB)和iPhone 17 Pro/ 17 Pro Max(12GB)。

也就是说,当时冲着AI名义买了iphone 16和iPhone 17标准版的用户,如今却被苹果背刺,被排除在顶配AI体验之外。

这才是苹果必须直面的问题:苹果的AI承诺正在不断消耗用户信任,品牌该如何兑现曾经对消费者的承诺?苹果的产品换机周期,与AI技术迭代速度,该如何和对用户的承诺保持一致?

苹果用下一代的AI想象力卖出当下的硬件,但现实追上了想象,硬件的门槛又提高到了下下一代。

库克还没有回答

「如果我是特纳斯,我会让库克吞下这一切,明年再做自己的盛大登场。」

这是美国科技论坛上流传的一种观点,虽偏向阴谋论,但也侧面印证了特纳斯缺席本届WWDC舞台的事实。传闻中,苹果新任CEO将主持今年9 月份的iPhone 18发布会,那才是他真正的舞台。

对于库克来说,AI之战是一场妥协的战斗。

外媒此前测算,苹果与谷歌签订多年授权协议,采用定制版Google Gemini模型(约1.2万亿参数)支撑新一代Siri的云端高阶功能,每年相关费用约10亿美元。当然,苹果也要支付另外的算力成本。

对比另一组数据:2026年 5月,SpaceX提交IPO申请文件,文件披露Anthropic与 xAI达成合作,每月出资12.5亿美元租用Colossus数据中心超22万张NVIDIA GPU,折合年成本达150亿美元。

因此在AI赛道上,库克选择了成本优先策略:不参与大模型军备竞赛,每年花费10亿美元租用谷歌模型,而非像Anthropic那样斥资150亿美元自研模型。现阶段,搭建前沿大模型的综合成本远高于租赁模式,差距十分明显。

如今苹果与谷歌联合打造了五款模型,分为本地端与云端两大类。其中本地设备模型之一AFM 3 Core,是30亿参数通用模型的全新升级版,综合能力大幅提升。

而AFM 3 Core Advanced是目前苹果最强的本地端模型,原生支持图文、语音等多种形式。它不仅语音交互更自然,语音识别精度也更高。该模型参数规模达200亿,采用稀疏架构,运行时会根据实际需求激活10亿至40亿参数,运行效率出色,且针对苹果旗舰芯片完成深度适配优化。

在私有云上,苹果也有三款模型,苹果承诺用户数据不会留存,也不会分享给任何人,包括苹果自身:

1、AFM 3 Cloud:云端主力模型,兼顾运行速度、效率与综合性能;

2、ADM 3 Cloud(图像模型):主打图片生成和编辑,可实现高阶修图、趣味图像创作等功能;

3、AFM 3 Cloud Pro:云端最强模型,专门应对高难度场景,比如复杂智能操作、深度逻辑推理等。

其中AFM 3 Cloud Pro部署在谷歌云的英伟达GPU集群上,也是谷歌参与程度最高的一款模型。

苹果用户调用AFM 3 Cloud Pro运行Siri功能时,苹果隐私保护体系依旧生效:数据在传输前完成脱敏处理,协议明确禁止谷歌将数据用于模型训练,但运算过程依托谷歌的基础设施完成。

这就好像苹果造了一栋AI大楼,楼里的保安是苹果的,摄像头是苹果的,门锁是苹果的。但盖楼用的钢筋,是谷歌的,而且顶层最贵的那套房,也是在谷歌的地基上盖的。

这栋楼,就是库克留给特纳斯的帝国遗产,它是一个现实,映射出苹果在AI层面的基础设施投入起步晚、节奏慢,但它同时也是一个阶段性成果,毕竟苹果已经有了AI之战的防守阵地,即便这个阵地是联合对手一起打造的。

库克任期内没有解决的问题,现在原封不动地转移到了特纳斯的桌上,这场AI时代的叙事刚刚开始,它要跟上世界的节奏,但书写它的主人,已经换了。

审校|陈秋霖

注:文/赵卫卫,文章来源:蓝洞商业(公众号ID:value_creation),本文为作者独立观点,不代表亿邦动力立场。

文章来源:蓝洞商业

广告
微信
朋友圈

这么好看,分享一下?

朋友圈 分享

APP内打开

+1
+1
微信好友 朋友圈 新浪微博 QQ空间
关闭
收藏成功
发送
/140 0