3.【宇树科技推出人形机器人App Store，用户可下载动作预设】12月13日，宇树科技宣布首发人形 机器人App Store，用户可在其中下载动作和预设，一键完成复杂操作。目前公测已上线搞笑动作、 扭扭舞、李小龙三大预设。（钛媒体） 1.【甲骨文：与OpenAI的合作安排"没有延迟"】当地时间12月12日，甲骨文公司否认为OpenAI建造 的数据中心完工时间从2027年推迟至2028年的报道。甲骨文发言人称，选址和交付时间表是在协议 签署后与OpenAI密切协调后确定的，并经双方共同同意。所有需要满足合同承诺的节点均未出现延 误，所有里程碑仍按计划推进。（界面新闻） 2.【谷歌称将Gemini的翻译功能引入谷歌的文本翻译】谷歌称将Gemini的翻译功能引入谷歌的文本 翻译，并推出耳机实时语音转语音翻译的测试版体验。（财联社） 4.【天桥脑科学研究院成立尖峰智能实验室，主攻类脑大模型和脉冲神经网络研发】12月13日，界面 新闻记者获悉，天桥脑科学研究院成立尖峰智能实验室（Spiking Intelligence Lab, SIL）。天桥脑 科学研究院创始人雒芊芊介绍，该实验室由中国科学院自动化所李国齐教授领 ...

Group 1 - The core viewpoint of the article highlights advancements in AI technology, particularly in visual models and robotics, showcasing the launch of the "Juzhou" model and the Booster T1 robot [2][3]. Group 2 - The "Juzhou" model, developed by Hunan Huishiwei Intelligent Technology Co., is the first domestically produced visual model based on pure domestic computing power, with the V1.5 version released on October 11, featuring enhanced performance and cross-platform capabilities from iOS to Android [2]. - The "Juzhou" model can generate 1024×1024 resolution images in seconds on iOS devices without internet access, boasting low cost, high quality, fast speed, and lightweight characteristics [2]. - The model's parameters have been reduced to 1/50, with training speed increased by 5 times and generation speed by 7 times, allowing it to become a specialized model for various industries [2]. - The Booster T1 robot, launched by Accelerated Evolution, is an upgraded version that can understand vague language commands and perform household chores autonomously [2]. - Perplexity CEO Srinivasan has transitioned from traditional investor presentations to using AI for investor roadshows, indicating a shift in how funding discussions are conducted [3]. - Elon Musk announced that Grok will soon have the capability to detect AI-generated videos and trace their online origins, addressing concerns over deepfake content [3].