Author: lethuphuong

📌 China’s UBTech company partnered with the humanoid robot center in Fangchenggang – a city bordering Vietnam – using the Walker S2 model launched in July. This is the world’s first robot capable of fully automatic self-battery-swapping in 3 minutes, carrying a load of 15 kg in each hand, with a speed of 7.2 km/h and hands capable of manipulation accuracy of less than 1 mm. Walker S2 is being piloted at border crossings to guide visitors, manage human traffic flow, assist with patrols, handle logistics, and support commerce; while also conducting inspections at steel, copper, and aluminum factories.

📌 Alibaba and ByteDance are moving AI training to Singapore–Malaysia to access Nvidia chips amid US export restrictions. Qwen and Doubao are becoming leading models; DeepSeek is the exception, training domestically thanks to accumulated Nvidia chips and collaboration with Huawei. Chinese enterprises also use Southeast Asian data centers to serve international customers.

📌 Generative AI like Claude can reduce the time required for an average task valued at $54–55 USD by 80% and could help increase US labor productivity by 1.8% per year over the next decade. When simulated for the entire US economy, labor productivity could increase by 1.8% annually, equivalent to a TFP increase of 1.1%/year. The largest productivity contributions come from programming (19%), followed by management (6%), marketing (5%), customer service (4%), and high school teachers (3%).

📌 From 2005–2024, China’s Tsinghua University filed 4,986 AI patents, surpassing the total number of patents from Harvard, MIT, Stanford, and Princeton combined; China accounts for over 50% of global AI patents. Tsinghua is likened to the Stanford + MIT + Carnegie Mellon of China. Tsinghua surpasses US universities due to its strong AI ecosystem, State support policies, and a STEM talent pool of up to 5 million people/year. The proportion of elite AI researchers in China (top 2%) increased from 10% (in 2019) to 26% (in 2022), while the US decreased from 35% to 28%. Many Tsinghua students choose…

📌 President Trump signed an Executive Order launching the Genesis Mission, “as urgent and ambitious as the Manhattan Project,” requiring the expansion of computing resources, increased access to federal scientific data, and the transition of AI to practical applications in science. This is a turning point as the US mobilizes federal data, supercomputers, and private partners to accelerate generative AI and foundation models. Within 90–270 days, data and infrastructure must be integrated, and AI must be deployed in advanced manufacturing, biotechnology, and nuclear energy sectors. The project inherits NAIRR, expands with new supercomputers from AMD, HPE, and Nvidia, strengthening the…

📌 A 526-page health report commissioned by the province of Newfoundland and Labrador, Canada, to Deloitte, valued at nearly $1.6 million USD, was found to contain errors suspected to be caused by generative AI, including incorrect citations and non-existent academic materials. Deloitte denied that AI wrote the report but admitted using AI to assist with citations. This is the second scandal after the Australian case, raising concerns about the accuracy and accountability when AI is used in government consultancy documents.

📌 The $100 million USD deal with Intuit marks a major milestone as OpenAI connects ChatGPT with the financial data of 100 million users for the first time via an opt-in mechanism, creating a generation of financial AI agents capable of forecasting, analysis, and tax processing. This move supports the goal of over $20 billion USD in annual revenue and pushes OpenAI into the enterprise market, while Intuit leverages AI to expand services and accelerate growth.

📌 Alibaba Cloud announced a partnership with AI Singapore (AISG) to launch the new Qwen-Sea-Lion-v4 version, built on Alibaba’s Qwen3-32B foundation model. This version marks a significant step for Singapore in developing a regional LLM, combining open-source AI and indigenous data. With the Qwen3-32B foundation and 100 billion Southeast Asian tokens, the model aims for indigenous, intelligent AI that deeply understands regional language and culture, reinforcing Singapore’s role as the AI hub of Southeast Asia.

📌 Google must double its AI capacity every 6 months to meet surging demand, aiming for a 1,000x increase in 4–5 years. AI infrastructure is the most expensive and competitive front. Google is increasing capacity not only by expanding infrastructure but also through more efficient models and custom chips. The goal is not to “spend the most” but to build an infrastructure that is “more stable, high-performance, and scalable” than competitors. 2025 capex is about $91–93 billion USD, cloud revenue increased by 34% to over $15 billion USD but still lacks compute capacity, with a backlog of $155 billion USD.…

📌 Many AI output evaluators become deeply skeptical after witnessing errors, biases, pressure for speed, and signs that businesses prioritize speed over safety. The rate of chatbots repeating false information increased to 35% in 08/2025, indicating the risk of widespread misinformation. Workers warn the public: AI is only as good as its input data, and the silent labor behind it is easily ignored. They call for questioning the data sources, ethics, and labor conditions to drive change.