![[b3616480-33cf-4fc1-8922-5c82f8205f4b.mp3]]

ChatGPT在macOS上线隐形智能笔记功能，支持自动转录和摘要

简报：

• OpenAI发布了专为会议记录、头脑风暴和个人知识管理设计的新笔记工具，目前独家上线于ChatGPT的macOS桌面应用。
• 用户可在应用内一键启动录音，工具会自动转录并提炼音频内容为结构化笔记，无需第三方机器人介入，保障会议私密性。
• 笔记直接保存在ChatGPT Canvas界面，支持智能生成关键点、行动项和总结性报告，还可通过提示词自定义输出。
• 该功能6月4日优先向Team用户开放，6月18日扩展至Pro、Enterprise和Edu用户，目前仅限macOS系统，其他平台尚未公布计划。
• 与近期升级的ChatGPT记忆功能结合，新笔记工具能参考用户历史记录，按用户习惯个性化生成内容；市场反馈普遍好评，但与Notion等竞品相比，协作性与多平台支持仍有差距。
• 除会议场景外，该功能还适用于整理讲座、制作学习材料、提炼研究内容等多元用途，并可结合API接入Slack、Google Docs等平台。

相关链接：

• AIbase: ChatGPT 新笔记工具正式上线
• AIbase: ChatGPT 笔记工具功能详解

ChatGPT Unveils Integrated Smart Notes on macOS, Featuring Automatic Transcription and Summarization

Briefing:

• OpenAI has launched a new note-taking tool designed specifically for meeting minutes, brainstorming, and personal knowledge management, currently exclusively available on the ChatGPT macOS desktop application. 📝
• Users can start recording with a single click within the app. The tool automatically transcribes and distills audio content into structured notes, without the need for third-party bots, ensuring meeting privacy. 🔒
• Notes are saved directly in the ChatGPT Canvas interface, supporting intelligent generation of key points, action items, and summary reports. Output can also be customized via prompts.
• This feature was initially rolled out to Team users on June 4th and will be extended to Pro, Enterprise, and Edu users on June 18th. Currently, it's limited to macOS, with no plans announced for other platforms.
• Combined with the recently upgraded ChatGPT Memory feature, the new note-taking tool can reference user history to personalize content based on user habits. Market feedback has been generally positive, but compared to competitors like Notion, there's still a gap in collaborative features and multi-platform support.
• Beyond meeting scenarios, this feature is also suitable for diverse uses such as organizing lectures, creating study materials, and extracting research content. It can also integrate with platforms like Slack and Google Docs via API. 💡

Related Links:

• AIbase: ChatGPT 新笔记工具正式上线
• AIbase: ChatGPT 笔记工具功能详解

特斯拉在德州奥斯汀正式试运营无人驾驶RoboTaxi，初期定价4.2美元

简报：

• 特斯拉于2025年6月22日在美国得州奥斯汀正式启动备受期待的RoboTaxi自动驾驶出租车服务，首批仅投入10辆Model Y汽车试营运，并配备真人驾驶员监督安全。
• 用户需通过专用叫车应用体验无人驾驶，首批抢先体验价格为每次4.20美元，乘客需年满18岁方可搭乘。
• 此举被视为特斯拉自动驾驶战略的重要里程碑，公司计划未来逐步扩展服务范围并上线无方向盘、无踏板的Cybercab车型。
• 本次启动也引发当地立法关注，部分得州议员联名建议特斯拉待自动驾驶新法生效后再推出服务。

相关链接：

• 36氪快讯报道
• 三立新闻网：特斯拉無人車22日試營運
• 三立新闻网：特斯拉Robotaxi專題
• 三立新闻网：特斯拉自駕黃金時代

Tesla Officially Pilot Tests RoboTaxi in Austin, Texas, with Initial Pricing at $4.20

Briefing:

• On June 22, 2025, Tesla officially launched its highly anticipated RoboTaxi autonomous ride-hailing service in Austin, Texas, USA. The initial pilot operation involves only 10 Model Y vehicles, each equipped with a human safety driver for supervision. 🚕
• Users must access the autonomous driving experience via a dedicated ride-hailing app. The first-wave, early-bird price is $4.20 per ride, and passengers must be at least 18 years old to board. 📲
• This move is considered a significant milestone in Tesla's autonomous driving strategy. The company plans to gradually expand service coverage and introduce the steering wheel-free, pedal-free Cybercab model in the future. 🚀
• This launch has also drawn attention from local legislators, with some Texas lawmakers jointly suggesting that Tesla should wait until new autonomous driving laws come into effect before fully rolling out the service.

Related Links:

• 36Kr News Flash Report
• SETN News: Tesla Driverless Cars Pilot Test on the 22nd
• SETN News: Tesla Robotaxi Special Report
• SETN News: The Golden Age of Tesla Self-Driving

OpenAI因商标诉讼被迫撤下Jony Ive相关合作信息

简报：

• 由于与来自iyO公司的商标纠纷，法院下达禁令要求OpenAI撤回所有涉及“io”名称的宣传材料。
• OpenAI已将包括与苹果前首席设计官Jony Ive（乔纳森·伊夫）合作相关的内容从官网的合作伙伴关系栏目及宣传视频中删除。
• OpenAI确认事件未影响收购Jony Ive和Sam Altman共同创立的设备初创公司io的交易进展，目前正在评估应对方案。

相关链接：

• OpenAI因商标纠纷问题将Nony Ive从网站的合作伙伴关系栏目删除
• OpenAI撤下与Jony Ive相关宣传材料：因商标纠纷遭法院禁令

OpenAI Forced to Remove Jony Ive Partnership Information Due to Trademark Lawsuit ⚖️

Briefing:

• A court injunction has ordered OpenAI to remove all promotional materials involving the "io" name, following a trademark dispute with a company named iyO.
• OpenAI has removed content related to its collaboration with former Apple Chief Design Officer Jony Ive from its official website's partnerships section and promotional videos. 🚫
• OpenAI confirmed that the incident has not affected the acquisition progress of io, a device startup co-founded by Jony Ive and Sam Altman, and is currently evaluating its response options. 🤔

Related Links:

• OpenAI Removes Jony Ive From Website Partnership Section Due to Trademark Dispute
• OpenAI Removes Jony Ive-Related Promotional Materials: Court Injunction Due to Trademark Dispute

MiniMax推出AI视频Agent工具，支持文本指令生成高清视频与人脸身份一致性

简报：

• MiniMax公司近日发布新型视频Agent工具，用户可通过一句文本描述自动生成最高720p、最长6秒的高清视频，并计划未来扩展至10秒；
• 工具支持上传人脸图片，实现视频中人物身份特征高度一致，适用于虚拟主播、品牌代言等个性化场景；
• MiniMax视频Agent依托多模态AI技术，兼具文本处理、图像生成和视频合成能力，并提供API接口便于企业和开发者集成；
• 当前工具在时长、速度和多语言支持方面仍有提升空间，但凭借易用性和精准人脸一致性，在市场上形成差异化竞争优势。

相关链接：

• MiniMax重磅发布视频Agent工具
• MiniMax推出Hailuo Video Agent开放Beta
• MiniMax Hailuo 02用户基准测试超谷歌Veo 3

MiniMax Unveils AI Video Agent Tool: Text-to-HD Video & Consistent Facial Identity

Briefing:

• MiniMax recently launched a new Video Agent tool, enabling users to automatically generate high-definition videos up to 720p and up to 6 seconds long from a single text description. Future plans include extending this to 10 seconds. 🚀
• The tool supports uploading facial images to ensure high consistency in character identity features within the video, ideal for personalized scenarios like virtual anchors and brand endorsements. 🤩
• MiniMax's Video Agent is powered by multimodal AI technology, combining text processing, image generation, and video synthesis capabilities, with API interfaces available for easy integration by businesses and developers.
• While the current tool still has room for improvement in terms of duration, speed, and multilingual support, its ease of use and precise facial identity consistency provide a differentiated competitive advantage in the market. ✨

Related Links:

• MiniMax Unveils Major Video Agent Tool Release
• MiniMax Launches Hailuo Video Agent Open Beta
• MiniMax Hailuo 02 User Benchmark Exceeds Google Veo 3

华为盘古大模型5.5发布，实现五大基础模型全面升级

简报：

• 华为在开发者大会HDC2025上发布盘古大模型5.5，对自然语言处理、计算机视觉、多模态、预测、科学计算五大基础模型进行了全面升级
• 盘古5.5引入了业界首创的triplet transformer三元组统一预训练架构，显著提升预测模型的精度和泛化能力
• 全新718B深度思考NLP模型增强知识推理、工具调用、数学等领域能力，Agent与快慢思考融合使推理效率提升8倍
• 新版视觉大模型拥有300亿参数，是当前业界最大的视觉模型，提升工业等复杂场景识别精度
• 华为还推出面向医疗、金融、政务、工业、汽车五大行业的专用自然语言大模型以增强各行业AI能力
• 盘古大模型已在气象、电力等领域深度应用，有效优化天气预测和能源生产流程，并赋能智能驾驶及多模态智能体训练

相关链接：

• 华为云发布盘古大模型5.5
• 盘古大模型五大基础模型升级

Huawei Pangu Model 5.5 Launched, Featuring Comprehensive Upgrades to Five Foundational Models

Briefing:

• Huawei unveiled Pangu Model 5.5 at the Developer Conference HDC2025, bringing comprehensive upgrades to its five foundational models: Natural Language Processing, Computer Vision, Multimodal, Prediction, and Scientific Computing.
• Pangu 5.5 introduces the industry's first triplet transformer unified pre-training architecture, significantly boosting the accuracy and generalization capabilities of its prediction models. 🚀
• A new 718B deep-thinking NLP model enhances capabilities in areas such as knowledge reasoning, tool invocation, and mathematics. The integration of Agent with fast and slow thinking boosts inference efficiency by 8 times.
• The new version of the Vision Foundation Model boasts 30 billion parameters, making it the largest vision model in the industry, and improves recognition accuracy in complex scenarios such as industrial applications. 💡
• Huawei also launched specialized Natural Language Foundation Models tailored for five major industries: healthcare, finance, government, manufacturing, and automotive, to bolster AI capabilities across these sectors.
• The Pangu Model has been deeply applied in fields such as meteorology and electricity, effectively optimizing weather forecasting and energy production processes, and empowering intelligent driving and multimodal intelligent agent training. ✨

Related Links:

• Huawei Cloud Releases Pangu Model 5.5
• Pangu Model: Five Foundational Model Upgrades