20240930 – Zhipu AI | TeleAI | Tsinghua University | Tencent | more
1. CogView3, the open-source image generation model by Zhipu AI, has rolled out! Its generation speed is 10 times faster than SDXL!
Recently, Zhipu AI has open-sourced its groundbreaking models, CogView3 and the upgraded CogView-3Plus-3B, revitalizing the text-to-image generation field. CogView3’s innovative cascade diffusion approach generates high-quality images more efficiently than current leading models, achieving a remarkable 77% win rate over SDXL while reducing inference time significantly. Meanwhile, CogView-3Plus-3B builds on this success, introducing advanced techniques and expansive resolution options, enhancing its versatility for both casual and professional users.
Original Chinese article: https://www.aibase.com/zh/news/12136
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fwww.aibase.com%2Fzh%2Fnews%2F12136
2. Breakthrough! TeleAI has completed the training of the first fully domestically produced model with 10,000 cards and 10,000 parameters, and has officially open-sourced TeleChat2-115B.
TeleAI has achieved a significant milestone by successfully training the first domestically produced trillion-parameter model, TeleChat2-115B, using indigenous technology. This marks a new phase in the development and innovation of large models, reinforcing China’s commitment to self-sufficiency in artificial intelligence. The model has already demonstrated exceptional performance in logic reasoning and has been officially open-sourced to foster further advancements in this technology.
Original Chinese article: https://mp.weixin.qq.com/s/SvgPiGl5xy_vPuqtq77Byg
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fmp.weixin.qq.com%2Fs%2FSvgPiGl5xy_vPuqtq77Byg
3. Tsinghua University and Tencent jointly launched the fully open-source multi-modal framework Oryx, supporting ultra-long video input.
A groundbreaking multi-modal large language model, ORYX, developed by researchers from Tsinghua University, Tencent, and Nanyang Technological University, is revolutionizing AI’s understanding of the visual world. ORYX excels in processing diverse visual inputs, utilizing advanced components like OryxViT to interpret images and videos across various resolutions, while its dynamic compression capabilities ensure efficient data handling. With outstanding performance in visual-language benchmarks, ORYX paves the way for enhanced human-computer interaction and has immense potential in fields such as intelligent monitoring and autonomous driving.
Original Chinese article: https://www.aibase.com/zh/news/12107
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fwww.aibase.com%2Fzh%2Fnews%2F12107
4. Turing Award winner Andrew Yao: In the future, the number of robots will surpass that of humans.
Turing Award winner Andrew Yao believes that in the future, robots will outnumber humans and become significant members of society. His insights underscore the growing importance of artificial intelligence and robotics in shaping our world. Yao’s perspective invites us to consider the implications of this technological evolution on our daily lives and social structures.
Original Chinese article: https://finance.sina.com.cn/tech/digi/2024-09-30/doc-incqwyyw5195377.shtml
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Ffinance.sina.com.cn%2Ftech%2Fdigi%2F2024-09-30%2Fdoc-incqwyyw5195377.shtml
5. Last year, China added more than 270,000 industrial robots, seven times more than the United States, accounting for half of the global total.
In 2023, China significantly outpaced the global industrial robotics market by adding over 276,000 new robots, which is seven times more than the United States and accounts for half of the world’s total installations. This growth underscores China’s dominance in the sector, driven by a comprehensive manufacturing ecosystem and a push for increased automation. As the country continues its technological advancement, the industrial robot market is expected to grow annually by 5-10% through 2027.
Original Chinese article: https://36kr.com/p/2970779073666949
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2F36kr.com%2Fp%2F2970779073666949
6. Breaking news! Huitun AI has launched the first AI voice model designed specifically for live streaming.
Huitun AI has unveiled an innovative AI voice model specifically designed for live streaming, distinguishing itself from traditional mechanical voice models with its natural-sounding human-like tones. This cutting-edge model boasts over a thousand unique voice styles, rich emotional expressions, and supports multiple languages and dialects, making it a versatile tool for content creators on platforms like TikTok and YouTube. The technology aims to enhance live broadcasting experiences with features like ambient sound integration and automated content moderation for safer streaming.
Original Chinese article: https://www.163.com/dy/article/JD3MMMB4055641DN.html
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fwww.163.com%2Fdy%2Farticle%2FJD3MMMB4055641DN.html
7. Tencent and Tongji University have reached a collaboration: jointly building the “Big Design and Big Model Initiative,” launching AI intelligent agents for university professors.
Tencent and Tongji University have partnered to establish the “Big Design and Big Model Initiative (TT),” aiming to create an innovative ecosystem for design models that bridge academia and industry. This initiative includes the launch of “Prof. Lou,” an AI designed to address various design-related inquiries, leveraging the expertise of university professors and extensive data. Together, they will explore advancements in AI technology across multiple design disciplines, enhancing educational and research opportunities.
Original Chinese article: https://finance.sina.com.cn/tech/digi/2024-09-27/doc-incqrmin7600046.shtml
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Ffinance.sina.com.cn%2Ftech%2Fdigi%2F2024-09-27%2Fdoc-incqrmin7600046.shtml
8. Let computing power serve the public like water and electricity; China Mobile’s computing network brain achieves large-scale commercial use.
China Mobile has launched its “Computing Network Brain” for large-scale commercial use, aiming to make computing resources as accessible as electricity. This innovative system centralizes management and intelligent scheduling of computing resources, facilitating industry digital upgrades and transforming computing into a societal utility that can be accessed effortlessly. The platform supports efficient resource allocation across various sectors, fostering innovation and reducing operational costs for businesses.
Original Chinese article: https://digital.gmw.cn/2024-09/28/content_37589304.htm
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fdigital.gmw.cn%2F2024-09%2F28%2Fcontent_37589304.htm
9. Zhipu API launches its biggest promotional event: One million users receive 100 million tokens for free to enjoy the most powerful model, GLM-4-Plus.
The Zhipu AI Open Platform bigmodel.cn has launched an exciting promotion for users during the National Day celebrations, offering 1 billion tokens to each of its million users, alongside a 10% to 50% discount on API calls. This initiative aims to showcase the capabilities of their flagship models, GLM-4-Plus and GLM-4V-Plus, which have gained recognition for breaking into the top tier of international language models. Users are encouraged to explore the potential of these advanced models on the bigmodel.cn platform.
Original Chinese article: https://mp.weixin.qq.com/s/7QZRoxQC94qxcjHH5Xn98w
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Fmp.weixin.qq.com%2Fs%2F7QZRoxQC94qxcjHH5Xn98w
10. Hisense AI TV E7N officially launched: Introducing the industry’s first twelve AI intelligent entities, leading a new benchmark for AI picture quality.
Hisense has officially launched the AI TV E7N, introducing an industry-first collection of twelve AI intelligent systems that set a new standard for AI image quality. Featuring cutting-edge technologies like the ULED AI picture quality platform and its proprietary Starsea AI model, this device promises enhanced viewing experiences with optimal detailing and deep learning capabilities. The E7N will be available in various sizes and price points, catering to diverse consumer needs in the premium market.
Original Chinese article: https://tech.china.com.cn/elec/20240930/405951.shtml
English translation via free online service: https://translate.google.com/translate?hl=en&sl=zh-CN&tl=en&u=https%3A%2F%2Ftech.china.com.cn%2Felec%2F20240930%2F405951.shtml
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.