China AI Native Industry Insights – 2024-11-18 – Douyin | Wechat | Quark | Kimi | more
1. Douyin is promoting its “V Project,” launching an AI avatar feature to enhance live stream interaction.
ByteDance Douyin has launched “Project V,” which introduces AI avatars designed to enhance live streaming engagement. These virtual characters can interact with users 24/7 using Doubao’s large language model. The project aims to deepen user engagement across multiple communication methods such as AI chat spaces and live streams.
Original Chinese article: https://www.aibase.com/zh/news/13295
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https://www.aibase.com/zh/news/13295
2. WeChat’s official account has quietly launched an AI voice cloning feature, but it hasn’t been fully rolled out yet.
Tencent WeChat has introduced a new AI voice cloning feature in version 2.29.1 of its public account platform. Users can create a personalized voice by recording a short sample and use it to read articles aloud. The feature, still in limited testing, aims to enhance user engagement and provide a more personalized experience.
Original Chinese article: https://www.aibase.com/zh/news/13298
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https://www.aibase.com/zh/news/13298
3. Quark’s New “Academic Search” AI Application Becomes Top Choice for Young People on PCs.
Quark has launched its “Academic Search” AI tool, which uses a database of over 100 million academic papers to help users efficiently retrieve information. Quark’s PC application ranked first in downloads for domestic AI PC applications in 2024, with over 50% of users under 25. Popular features include academic outline generation and PPT creation.
Original Chinese article: https://www.aibase.com/zh/news/13296
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https://www.aibase.com/zh/news/13296
4. Benchmarking O1: Kimi Releases China’s Most Competitive AI Model with Enhanced Mathematical Reasoning.
Moonshot AI’s new k0-math model has achieved performance comparable to OpenAI’s o1 series, outscoring o1-mini and o1-preview in basic math benchmarks (93.8 vs 90 and 85.5). The model features enhanced reasoning capabilities and deliberate problem-solving approaches, while Kimi’s platform has reached 36 million monthly active users.
Original Chinese article: https://mp.weixin.qq.com/s/2_K6KFwqe7gE2sv5DIvmJQ
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https://mp.weixin.qq.com/s/2_K6KFwqe7gE2sv5DIvmJQ
5. Kunlun Tech’s Skywork 4.0 O1 Large Language Model to Begin Invitation-Only Testing on November 27th.
Kunlun Tech is launching Skywork O1 (Model 4.0), China’s first AI model with enhanced logical reasoning capabilities. The model features built-in thinking, planning, and reflection abilities, similar to OpenAI’s O1 approach. It will be released in two versions: an open-source version based on Llama 3.1 8B and an advanced version for the Skywork platform. This release represents a significant step in Chinese AI development, building upon the company’s previous Skywork 3.0 MoE model launched in April 2024.
Original Chinese article: https://mp.weixin.qq.com/s/OzSfOg7BGx6bSAL4zh72NQ
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https://mp.weixin.qq.com/s/OzSfOg7BGx6bSAL4zh72NQ
6. AIsphere’s PixVerse Opens Global Access to Text-to-Video and Image-to-Video Services.
AIsphere has announced the global launch of its PixVerse platform, offering text-to-video and image-to-video generation services. The platform, PixVerse V3, boasts significant improvements in underlying model capabilities, including multi-modal generation features like creative templates, lip-sync, story continuation, and style transfer. Users can create videos with detailed emotional expression and cinematic techniques, and the platform offers various creative templates to simplify the process. Access is now open globally.
Original Chinese article: https://www.aibase.com/zh/news/13280
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https://www.aibase.com/zh/news/13280
7. StepFun Technology: The iPhone 16 camera control button now supports calling Yuewen’s “Photo Ask” feature.
StepFun Technology Co., Ltd. announced that its “PhotoAsk” visual search function is now directly accessible via the iPhone 16’s camera control button. Users can activate PhotoAsk to solve problems visually, like calculating calorie intake from a meal photo or retrieving information from an image. Other Apple users can access the feature through the control center or lock screen widgets after updating to iOS 18.
Original Chinese article: https://www.aibase.com/zh/news/13302
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https://www.aibase.com/zh/news/13302
8. Rokid Glasses launched: Lightweight AR glasses priced at 2499 yuan, supporting AI translation and object recognition.
Rokid has launched its new AR glasses, Rokid Glasses, priced at 2,499 yuan. These glasses support customizable lenses and feature a convenient snap-on design. Powered by Alibaba’s Tongyi Qianwen multimodal model, they offer features including calling, search, object recognition, translation, navigation, and photo-based question answering. Rokid Glasses are expected to launch in the second quarter of 2025.
Original Chinese article: https://www.aibase.com/zh/news/13304
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https://www.aibase.com/zh/news/13304
——————————————————————————————————————
That’s all for today’s China AI Native Industry Insights. Join us at AI Native Foundation Membership Dashboard for the latest insights on AI Native, or follow our linkedin account at AI Native Foundation and our twitter account at AINativeF.