Singapore Tech News, Features & Reviews

China now surpasses the United States in consumer adoption of artificial intelligence. This trend is driven by Chinese technology organisations integrating AI directly into popular daily applications. American models may lead in capability, but China's approach to distribution through widespread platforms tailored to local habits provides a distinct advantage, according to a recent report by Morgan Stanley.

Meta is reportedly preparing to launch a new version of its smart glasses, which will not feature an integrated camera system. The device, internally known by the codename Luna, is anticipated for release this autumn. This development follows increased examination of smart glasses regarding their recording functions, according to a report from The Information.

Anthropic will open a Singapore office in October, making the city-state its fifth Asia-Pacific location as demand for its Claude AI assistant grows among businesses and public-sector organisations. The company says Singapore ranks second among 121 countries for Claude.ai usage relative to population, at 5.81 times the level expected for its population size. New ASEAN general manager Dale Finlay will lead customer and partner engagement from Singapore as the company hires loc

Beijing will prohibit most people and organisations from flying, possessing or storing drones and key components across the city from 15 November, under revised rules that designate the entire municipality as controlled airspace. The measures also restrict drones being transported into the capital and give existing owners a short window to sell, scrap or move their aircraft elsewhere, marking one of China’s strictest urban drone regimes ahead of tighter enforcement.

Unitree has introduced the G1+, a major hardware and interaction upgrade to its compact G1 humanoid robot, adding a more flexible neck, stronger and cooler motors, wider visual sensing, touch controls and improved audio. Pandaily reports a tax-inclusive Chinese price of about RMB95,000, while Unitree’s international page lists the standard model at US$15,000 before tax and shipping, positioning it as a relatively accessible research platform for developers.

Researchers from prominent Chinese universities and Big Tech are concentrating on a new objective in the artificial intelligence competition against the United States. This involves creating systems capable of generating enhanced versions of themselves without human intervention. A recent paper describes a five stage framework for recursive self improvement, termed RSI.

The European Union is preparing a wide-ranging legislative proposal. This plan restricts individuals under 15 years old from social media platforms, video-sharing services, artificial intelligence chatbots, and online games. This represents the most extensive regulatory effort to date for minors in the digital sphere, according to a European Commission document seen by Reuters.

Chinese researchers have developed a method to significantly enhance the durability of an emerging memory chip technology. This advancement addresses a key reliability concern, potentially facilitating its use in high-performance computing and future artificial intelligence systems. The global demand for more advanced semiconductors continues to grow, driving such research efforts.

Anthropic has informed its shareholders that the artificial intelligence organisation anticipates its adjusted operating income will be positive for a second consecutive fiscal quarter. This financial performance indicates a period of sustained operational efficiency and revenue generation for the developer of advanced AI models. This information was reported by The Financial Times, which cited multiple individuals possessing direct knowledge of the matter.

DeepSeek has released V4.1 Flash, a multimodal AI model designed to cut inference and memory costs while competing with larger systems on coding, cybersecurity and agent benchmarks. The Chinese developer says its new Causal Encoder-Decoder architecture activates only 8 billion parameters when processing input and 16 billion when generating output, despite using a 552 billion-parameter mixture-of-experts backbone and supporting contexts of up to one million tokens.