智东西美国圣何塞6月12日现场报道，今日，年度AI盛会AMD Advancing AI大会火爆开幕，全球第二大AI芯片供应商AMD亮出其史上最强AI 新品阵容——旗舰数据中心AI芯片、AI软件栈、AI机架级基础设施、AI网卡与DPU，全面展露与英伟达掰手腕的雄心宏图。 xAI技术团队成员Xiao Sun、Meta工程副总裁Yee Jiun Song、甲骨文云基础设施执行副总裁Mahesh Thiagarajan、HUMAIN CEO Tareq Amin相继 登台，与AMD董事会主席及首席执行官苏姿丰（Lisa Su）博士对谈。 OpenAI联合创始人兼CEO Sam Altman作为惊喜嘉宾压轴登场，透露OpenAI团队在MI300X和MI450上开展了一些工作。他评价说，MI450的内 存架构已为推理做好准备，相信它也会成为出色的训练选择。 此次AMD最新发布或预览的重点产品包括： 1、数据中心AI芯片AMD Instinct MI350系列：采用3nm制程，集成了1850亿颗晶体管，基于AMD CDNA 4架构，搭载288GB HBM3e内存， 内存带宽达到8TB/s，单GPU可运行5200个参数的大模 ...

Core Viewpoint - The rapid increase in pre-orders for Rokid Glasses highlights the growing interest and potential in the AI glasses market, driven by advancements in AI and AR technologies [1][2][3]. Company Overview - Rokid has received over 250,000 pre-orders for its AI glasses, which are set to launch at the end of June [1]. - The company is focused on refining its products and enhancing capabilities in response to high consumer expectations following its recent popularity [3][4]. Product Features - Rokid Glasses integrate AR display, camera, microphone, and speaker functionalities, distinguishing them from other market offerings that may lack comprehensive features [3][4]. - The glasses include advanced features such as AI-driven smart prompting and automatic page turning, enhancing user experience [6]. Technology Development - Rokid emphasizes self-research and development in key areas such as operating systems, voice interaction, and optical design, while collaborating with partners for supply chain aspects [7]. - The company has established AI and AR labs to focus on the development of these technologies, with a belief in the significant potential of AR glasses [9]. Market Dynamics - The AI glasses industry is experiencing a surge in competition, referred to as the "hundred glasses war," with many new players entering the market due to the maturity of generative AI models [10][11]. - The market is still in its early growth phase, with consumer demand for better products and experiences driving competition [12]. Global Expansion - Rokid is expanding its presence in over 80 countries, including key markets like Germany, the UK, the US, South Korea, and Japan, while preparing for localization and compliance in these regions [13]. - The company plans to open its capabilities to developers, fostering an ecosystem for creating scene-specific applications [14]. Future Trends - The trend towards lightweight AI glasses is expected to continue, with Rokid aiming to produce glasses weighing less than 50 grams, catering to a broader audience and enhancing daily usability [14][15].

1. 迁得快：让三方框架模型 "零成本"迁移，避免重复造轮子，同时模型精度完全对齐。 2 . 部署快：训转推全流程自动化，让大模型部署像执行一行命令一般敏捷高效 。 Figure 1 MindSpore生态快速迁移解决方案的技术架构 接下来，我们将揭开昇思 MindSpore的破局之道。 一、支持训练Day0迁移，构建跨框架的"无感智能翻译"能力 当大模型架构日新月异，开发者最怕被生态绑定。昇思 MindSpore通过三重兼容术打通主流 技术栈 ， 支持 主流 加速库 模型0代码迁移 ，通过精度自动对比工具 实现跨框架、跨版本、 跨策略快速调优， 精度对齐原模型， 实现 在分布式并行策略保持不变的情况下，训练性能提 升5 %+ 。 在训练生态方面，通过 MindSpeed /Megatron桥接层实现 PyTorch 模型零代码迁移，训练 脚本可直接运行； 通过动态图能力重构，昇思让 PyTorch 开发者获得"原生体验"，同时借力 MSAdapter 工具自动转换9 5 %以上接口， 主流模型如 DeepSeek 、 Pan g u 等 迁移损耗 逼近于零。 大模型发展日新月异，新的大模型层出不穷，参数规模 ...

媒体报道，DeepSeek今年1月推出的低成本、功能强大的人工智能模型震惊了全球科技行业后，阿里 巴巴集团的工程师们取消了假期，在春节假期期间继续工作，奋起直追，连夜加班赶进度。 阿里巴巴董事会主席蔡崇信（Joe Tsai）周三在巴黎举行的VivaTech科技大会上表示，中国充满活力的 消费互联网环境，加上本地工程师之间激烈的竞争文化，正不断推动中国在AI领域的创新。 他透露，阿里巴巴是在DeepSeek发布R1模型之后，才意识到自己在AI领域已经落后了。他举例说明了 中国科技行业竞争的激烈程度： "我们看了那篇论文，心想：'天啊，怎么我们落后了？我们之前也在做这些事情啊，'结果就是，我们 的工程主管决定说：'取消春节假期，所有人都留在公司，加班睡办公室，我们要加速开发进度。'几周 之内，我们就推出了自己的版本，也就是Qwen系列模型。它非常具有竞争力"。 作为与马云共同创办阿里巴巴的创始成员之一，尽管蔡崇信近来开始更加坦率地谈论公司所经历的困 难，但同时也表达了对未来发展的乐观态度。上个月在澳门举办的一场科技大会上，他曾提到阿里巴巴 经历了一系列挫折，但也强调公司"正走在一条非常好的路上"。 ⭐星标华尔街见 ...

"我们看了那篇论文，心想：'天啊，怎么我们落后了？我们之前也在做这些事情啊，'结果就是，我们的工程主管决定 说：'取消春节假期，所有人都留在公司，加班睡办公室，我们要加速开发进度。'几周之内，我们就推出了自己的版本， 也就是Qwen系列模型。它非常具有竞争力"。 硬·AI 作者 | 赵雨荷 编辑 | 硬 AI 媒体报道，DeepSeek今年1月推出的低成本、功能强大的人工智能模型震惊了全球科技行业后，阿里巴巴 集团的工程师们取消了假期，在春节假期期间继续工作，奋起直追，连夜加班赶进度。 阿里巴巴董事会主席蔡崇信（Joe Tsai）周三在巴黎举行的VivaTech科技大会上表示，中国充满活力的消 费互联网环境，加上本地工程师之间激烈的竞争文化，正不断推动中国在AI领域的创新。 他透露，阿里巴巴是在DeepSeek发布R1模型之后，才意识到自己在AI领域已经落后了。他举例说明了中 国科技行业竞争的激烈程度： 蔡崇信还说，开源大型语言模型（LLMs）将带动人工智能应用的激增，并推动云计算需求的增长。目 前，阿里巴巴正经历"一个巨大考验的时期"之后，进一步聚焦自身庞大业务帝国的重心。 他说，阿里巴巴选择将Qwen系列 ...

Group 1 - The A-share market showed a mixed performance with the Shanghai Composite Index up by 0.07%, Shenzhen Component Index up by 0.05%, and the ChiNext Index up by 0.5%, while the North Star 50 and Sci-Tech 50 indices fell by 0.38% and 0.12% respectively, with a half-day trading volume of 793.75 billion yuan [1] - The IP economy is gaining traction, with companies that have strong IP product development and operation capabilities expected to benefit from increased IP traffic, as highlighted by Guotai Junan's research report [1] - The media sector is experiencing a surge in AI applications, with 2023 anticipated to be a pivotal year for the explosion of open-source large models in China, reshaping the application landscape [1] Group 2 - The Entertainment and Media ETF (516190) closely tracks the China Securities Entertainment and Media Index, covering emerging sectors such as gaming, advertising, film and animation, tourism, and digital publishing, with nearly 70% of its mainboard components and over 30% from the ChiNext board, showcasing both stable performance and high growth potential [2] - The entertainment and media sector is driven by multiple factors including policy support, technological innovation, and cultural export, presenting significant investment value [2]

Group 1 - Meta is establishing a new lab focused on "Superintelligence" to develop AI systems that surpass human intelligence in reasoning, problem-solving, creativity, and decision-making [1][3] - Meta has agreed to acquire 49% of Scale AI for $14.8 billion, which is approximately 106.14 billion RMB [1][3] - Alexander Wang, the 28-year-old CEO of Scale AI, is invited to join Meta's new lab, highlighting Meta's strategy to attract top talent in the AI field [1][4] Group 2 - Meta is offering compensation packages ranging from seven to nine figures to recruit top researchers from companies like OpenAI and Google, with some already agreeing to join [4][9] - Scale AI, founded in 2016, provides data labeling solutions and reported a revenue of $870 million in the previous year, with expectations to double to over $2 billion this year [3][9] - Meta's AI efforts are led by two groups: a generative AI team and a fundamental AI research lab, with Yann LeCun, a Turing Award winner, overseeing the latter [4][9] Group 3 - Meta's recent AI model testing faced criticism, with external researchers questioning the objectivity of its benchmark tests [5][8] - The company aims to regain its competitive edge in AI, especially after the rise of ChatGPT, which has intensified competition in the tech industry [9][10] - Meta's previous focus on open-source large models and social platform AI tools has led to a fragmented strategy, prompting the need for a more cohesive approach [10]

一、突传出售：明星项目的无声"叛逃" 就在这几天，开源社区突传爆炸性消息： 49.8K Star 的网盘聚合工具 Alist 已被整体出售 。 事件导火索源于用户发现其官方域名从 alist.nn.ci 悄然更换为 alistgo.com ，中文文档被大 规模修改，新增商业化内容（如 VIP 技术支持 QQ 群、付费定价策略），而原开发者 Xhofe 突然从 所有社群消失且长期沉默。 更令用户警觉的是，新提交的代码中出现 设备数据收集模块 （后因社区抗议被紧急撤回），桌面版 下载链接甚至指向腾讯云 COS 存储（因侵权遭封禁）。这些异常操作与开源透明度背道而驰，迅速 点燃社区质疑。 二、原开发者回应：沉默后的"有限承诺" 面对社区声浪，原开发者 Xhofe 于 6 月 11 日首次回应，但声明内容引发更大争议： 1. 未否认出售事实 ，仅模糊承认"移交公司运营"； 2. 未披露交易细节 ，收购方身份、金额、用户数据归属均未说明； 3. 审查权限存疑 ：所谓"帮忙审查"是否具备实际约束力遭强烈质疑，甚至还有网友调侃："不会连 账号都卖了吧？" 社区担忧，创始人此举实则为商业化让路，而"分支保护"承诺难以抵挡资本 ...

证券代码：301345 证券简称：涛涛车业 浙江涛涛车业股份有限公司 3、除美国市场外，公司其他市场的销售情况如何？ 答：公司非美市场的销售总体呈现增长趋势，主要系全地形车 销售在继续扩大。公司将通过加大研发力度、积极拓展销售渠道等 措施，争取在非美地区实现更多品类产品的销售。 4、随着技术的不断发展，预计机器人行业发展前景和市场需求 巨大，请问公司如何看待或会考虑进入吗？ 投资者关系 活动主要内 容介绍 答：一直以来，公司对于前沿技术的发展都较为关注和跟踪， 基于目前全球科技革命与产业变革的浪潮之巅，各种机器人产品的 市场需求潜力无限。在持续做好主业、不影响自身经营情况的同时， 公司经过较长时间的考虑和衡量，拟适时适度的进入机器人行业， 并视实际进展情况做进一步的规划。日前，公司的美国子公司 RevEdge Inc.已与美国 DPSH SYNDICATE .（也称"K-Scale Labs"） 签订投资协议，主要情况为：①K-Scale Labs 系美国一家致力于打 造开源机器人 AI 平台的公司，核心团队来自于行业顶尖阵营，获得 多家硅谷顶级资本的强力加持。作为人形机器人领域的创新玩家， 其核心技术优势在 ...

Group 1 - OpenAI has postponed the release of its anticipated open-source model to "later this summer" instead of June, as announced by CEO Sam Altman [1] - The open-source model aims to match the complex reasoning capabilities of GPT-4o and surpass leading open-source models like DeepSeek's R1 [2] - The AI market competition is intensifying, with new models being launched by competitors such as Mistral and Qwen, which are capable of switching between deep reasoning and traditional quick responses [2] Group 2 - Altman acknowledged that OpenAI has historically made mistakes in its open-source strategy, and the new model is seen as a crucial step to repair developer relations [2] - There are speculations that the delay may be a strategic move to counter DeepSeek's upcoming R2 model, which is expected to be released soon [2][3] - DeepSeek R2 is anticipated to have significant upgrades in technical architecture, functionality, and resource efficiency, with a predicted 87% reduction in AI invocation costs [3] Group 3 - DeepSeek's founder, Liang Wenfeng, emphasizes the goal of making China a contributor to innovation rather than a passive participant [4] - DeepSeek's product iteration schedule is robust, with plans for major updates every quarter, including the upcoming V2.5 and V3 versions [4]