目录
OpenAI联手前苹果设计师,欲推AI硬件“Sweetpea”挑战AirPods
OpenAI Teams Up with Former Apple Designer to Launch "Sweetpea" AI Hardware, Challenging AirPods
Anthropic Launches AI Programming Assistant Cowork 🚀
1X Unveils New AI Model: Robots Learn Tasks by Watching Videos 🤖
Google AI Shopping Feature Sparks 'Price Gouging' Controversy, Company Denies Claims 🧐
Google Unveils Two New Medical AI Models, Enhancing Language and Speech Capabilities 🚀
单词学习:![[df10db74-f55b-4cfe-b864-56000f72940d.mp3]]
新闻播报:![[8e9a4f51-1624-4a04-af0e-0fb880b616fb.mp3]]

OpenAI联手前苹果设计师,欲推AI硬件“Sweetpea”挑战AirPods
简报:
- 消息称,OpenAI正与前苹果首席设计官Jony Ive合作,秘密开发一款代号为“Sweetpea”的AI音频设备,旨在对标苹果AirPods。
- 该设备主体为鹅卵石状金属材质,包含两个可取出并佩戴于耳后的胶囊状组件,工业设计独特。
- 硬件预计将搭载2纳米工艺的智能手机级芯片,并已要求富士康为生产做准备。
- “Sweetpea”是OpenAI计划在2028年前推出的五款硬件中的首发项目,最早可能于今年9月发布,首年产量目标为4000万至5000万台。
相关链接:
Briefing:
- Reports indicate that OpenAI is secretly developing an AI audio device, codenamed "Sweetpea," in collaboration with former Apple Chief Design Officer Jony Ive. The device aims to rival Apple AirPods. 🤫
- The main body of the device is described as pebble-shaped metal, featuring two detachable, capsule-like components designed to be worn behind the ear, boasting a unique industrial design. ✨
- The hardware is expected to be powered by a smartphone-grade 2nm process chip, and Foxconn has reportedly been asked to prepare for its production.
- "Sweetpea" is the first of five hardware projects OpenAI plans to launch before 2028. It could be released as early as September this year, with a first-year production target of 40 to 50 million units. 🚀
Related Links:
Anthropic推出AI编程助手Cowork
简报:
- AI安全与研究公司Anthropic宣布,正式发布其新款AI编程助手,命名为“Cowork”。
相关链接:
Anthropic Launches AI Programming Assistant Cowork 🚀
News Brief:
- Anthropic, an AI safety and research company, has officially announced the launch of its new AI programming assistant, named "Cowork". 🤖✨
Related Links:
research
/rɪˈsɜːrtʃ/
n. 研究;v. 研究
▶ "He is doing research in artificial intelligence."
[例句] 他正在进行人工智能方面的研究。
1X推出新AI模型,机器人看视频即可自学任务
简报:
- 由OpenAI投资的机器人公司1X发布了一款新的AI“世界模型”,旨在提升其人形机器人Neo的学习能力。
- 该模型使Neo机器人能够通过观看人类执行任务的视频,自主学习并模仿复杂的、多步骤的物理操作。
- 这种“视频到动作”(video-to-action)的数据驱动方法,使机器人能从非结构化视频中学习,大幅减少对人工编程的依赖,加快了新技能的获取速度。
相关链接:
1X Unveils New AI Model: Robots Learn Tasks by Watching Videos 🤖
Brief:
- 1X Technologies, a robotics company backed by OpenAI, has launched a new AI "world model" designed to enhance the learning capabilities of its humanoid robot, Neo.
- This model enables Neo robots to autonomously learn and imitate complex, multi-step physical operations simply by watching videos of humans performing tasks. 📺
- This data-driven "video-to-action" approach allows robots to learn from unstructured videos, significantly reducing reliance on manual programming and accelerating the acquisition of new skills. ✨
Related Links:
谷歌AI购物新功能引“宰客”争议,官方否认
简报:
- 谷歌为其AI购物智能体推出全新通用商务协议(UCP)后,一家消费者经济监督机构发出警告。
- 该机构指出,协议中的“个性化追加销售”功能可能被用于分析用户聊天数据,并以此向其收取更高费用,即“宰客”。
- 谷歌官方回应称指控不实,并解释“追加销售”是零售商展示高端产品选项的标准营销方式,用户拥有最终决定权。
- 谷歌强调,公司政策严格禁止商家在谷歌平台展示高于其自有网站的价格,且名为“直接优惠”的试点功能旨在提供更低价格。
相关链接:
Brief:
- A consumer economic watchdog agency has issued a warning after Google launched a new Universal Commerce Protocol (UCP) for its AI shopping agent.
- The agency highlighted concerns that the "personalized upselling" feature within the protocol could potentially be used to analyze user chat data and subsequently charge them higher prices—effectively "price gouging." 💸
- Google officially responded, stating the accusations are false. The company clarified that "upselling" is a standard marketing practice for retailers to present premium product options, emphasizing that users retain the final decision.
- Google stressed that its company policy strictly prohibits merchants from displaying prices on Google's platform that are higher than those on their own websites. Furthermore, a pilot feature named "Direct Deals" aims specifically to offer lower prices. 🛍️
Related Links:
谷歌推出两款医疗AI新模型,强化语言与语音能力
简报:
- 谷歌发布了其最新的医疗AI模型MedGemma 1.5。
- 同时,谷歌还推出了一款专用于医疗领域的语音转写模型MedASR。
相关链接:
Google Unveils Two New Medical AI Models, Enhancing Language and Speech Capabilities 🚀
Brief:
- Google has released its latest medical AI model, MedGemma 1.5. 🤖
- Additionally, Google also launched MedASR, a speech transcription model specifically designed for the medical domain. 🗣️
Related Links:
Transcription
/trænˈskrɪpʃn/
n. 抄写;转录
▶ "The transcription of the interview took many hours to complete."
[例句] 访谈录音的转录花了很多时间才完成。
本文作者:topwind
本文链接:
版权声明:本博客所有文章除特别声明外,均采用 BY-NC-SA
许可协议。转载请注明出处!