Advertisement
Advertisement
Vincent Chow
Reporter, Tech
Vincent Chow is a technology reporter covering AI, with a focus on how society navigates the emergence of increasingly powerful AI systems. He previously covered Chinese society and was awarded a SOPA award for his culture reporting. He is currently supported by the Tarbell Center for AI Journalism.
Vincent Chow is a technology reporter covering AI, with a focus on how society navigates the emergence of increasingly powerful AI systems. He previously covered Chinese society and was awarded a SOPA award for his culture reporting. He is currently supported by the Tarbell Center for AI Journalism.
Areas of Expertise:
AI, AI safety, Chinese societyLanguages Spoken:
Cantonese, English, Mandarin
Dubbed one of China’s new ‘AI tigers’, Zhipu could become one of the first major large language model start-ups globally to go public.
The Chinese AI pioneer also releases first robot dogs under new venture Ace Robotics.
Manus 1.5, launched in October, boasts improvements in speed and reliability across research and web development tasks.
The report calls on US actors to engage with Chinese firms as the AI landscape continues to evolve.
Advertisement
US start-up Gloo’s Flourishing AI Christian benchmark tested how the output of AI models ‘help people flourish’, CEO Pat Gelsinger said.
Washington’s evaluation highlights an increasing number of Chinese companies developing world-leading open-weight AI models, not just DeepSeek.
The turning point came when DeepSeek released its V3 and R1 models, which were on par with OpenAI’s GPT and Meta’s Llama models at the time.
Throughout 2023 and 2024, Chinese firms keen to catch up with the US used Meta’s Llama to help them build their own models.
Reported move to reinvigorate Meta’s faltering AI effort with new ‘Avocado’ model is another likely win for Chinese AI.
SenseTime may have slipped from the spotlight as one of China’s “AI dragons”, but its tilt at multimodal, real-world AI hints at a comeback.
Just two weeks into its public beta, Qwen has shown explosive growth, underscoring Alibaba’s challenge to global rivals OpenAI and Google.
DeepSeek’s V3.2-Speciale equals Google’s Gemini 3 Pro in reasoning capabilities, the Chinese start-up says.
Donations surpass amount collected in first week after 2008’s devastating Sichuan earthquake.
Government will do its utmost to handle aftermath and resettle residents affected by the fire, Financial Secretary Paul Chan says in a blog post.
‘Imagine owning the brain of one of the best mathematicians in the world for free,’ Hugging Face CEO Clement Delangue says in a post on X.
The research could help reduce training and inference costs for Alibaba’s next generation Qwen models without sacrificing accuracy.
China’s AI regulatory framework is taking shape as its open-source systems gain popularity worldwide, according to paper.
As death toll rises, residents offer condolences to victims and call for support and blood donations.
Pouring money into increasing computation resources and training data can still yield better AI models, Yao Shunyu says.
OpenAI’s new ‘shopping research’ tool resembles popular AI shopping services on Chinese e-commerce platforms like Tmall and Taobao.
By publishing training data and pipelines, the Allen Institute for AI hopes its Olmo models can win trust – and chip away at China’s lead.
LingGuang, an AI app that generates custom software applications, sees a surge in popularity.
It highlights the daunting challenge faced by Chinese AI start-ups to compete against firms like Google with massive resources.
Qwen has the potential to become super-app, the same way Tencent’s WeChat defined the mobile internet age, analysts say.
The new pricing strategy reflects heightened competition in China’s foundational model market.
Tencent is developing physics-aware ‘world models’ to advance spatial intelligence, joining global players such as DeepMind and xAI.
Amid chip shortages, China’s AI start-ups are re-engineering their algorithms, hoping more efficient architecture will do the heavy lifting.
In a Reddit forum, a Moonshot AI executive says start-up is ‘outnumbered’ by US rivals in terms of ‘high-end GPUs’ used for training models.
Remarks by CEO Liang Wenfeng’s delegate highlight potential dangers of race to build superintelligent systems.
