Moonshot AI, the Chinese startup valued at $35 billion, is pushing into international markets with a new sales role posted in Singapore. The company, backed by Alibaba, Tencent, Meituan, and China Mobile, is also preparing for a possible Hong Kong IPO by dismantling its Cayman Islands structure. Moonshot's Kimi K3 model, a 2.8-trillion-parameter system, positions it against frontier releases from OpenAI and Anthropic.
Mistral AI is entering the speech and voice agent space with Voxtral 3, a family of models that includes three transcription models, one text-to-speech model, and a voice agent SDK built on LiveKit. The models support 50 languages, diarization, and context biasing. Pricing starts at $0.0035 per minute for batch transcription and $13 per million characters for speech output.
Meta and Google are racing to release the first widely adopted AI digital assistant. Google brings its AI models and broad product reach, while Meta leverages social networks that already connect with billions of people. Both face a risk: AI agents could reduce advertising revenue by pulling apps and websites out of the user experience.
New York City has proposed rules requiring businesses to disclose when customers are interacting with AI rather than a human. Companies must also provide a straightforward path to reach a real person. Fines start at $525 for violations. The rules would apply citywide, with exemptions for airlines and national banks.
A Sopra Steria Next report finds that fewer than 20% of AI-influenced decisions have an identified owner or clear accountability. Fewer than 25% of AI decisions in production can be traced and replayed from start to finish. The report calls for a five-layer framework covering data, models, platforms, APIs, and decisions to improve governance.
Espressif Systems launched the ESP-Mosaico, an ESP32-S31 development platform aimed at edge voice, multimedia, and motion sensing applications. The platform features a 480 by 480 OLED touchscreen, haptic feedback, built-in battery, and USB Type-C. It supports WiFi 6, Bluetooth 5.4, and Zigbee or Thread connectivity, and includes a Camera Module with a 3.2 MP sensor.
Central bankers are increasingly focused on AI's impact on jobs and wages. Donald Kohn, a former Federal Reserve vice chairman, said workers feel insecure about their jobs because of AI. Many Asian economies, including China and Indonesia, face risk if AI demand slows. The IMF and World Bank plan to discuss the topic at upcoming meetings in Basel.
A letter to the editor urges policymakers to slow advanced AI deployment until regulations and safety frameworks are in place. The author warns about risks to jobs, privacy, and democratic processes from deepfakes, and highlights heavy energy and water use by data centers as an environmental concern.
TAITRONICS 2026 opened in Taipei on October 25, connecting Taiwan's US$192 billion AI hardware supply chain with nearshoring hubs like Mexico. The trade show promotes collaboration among industry players exploring AI hardware, software, and solutions.
AI app creation is raising copyright questions for tech firms. People are using AI to build apps based on copyrighted ideas, drawing comparisons to past battles like browser wars, ad blockers, and Napster. The author expects a mix of legal pressure, agreements, and user-created tools before tech firms fully adapt.
Key Takeaways
- Moonshot AI, valued at $35 billion, is expanding into Singapore with a new sales role and preparing for a possible Hong Kong IPO.
- Moonshot's Kimi K3 is a 2.8-trillion-parameter model that competes with frontier releases from OpenAI and Anthropic.
- Mistral AI launched Voxtral 3, a speech model family and voice agent SDK supporting 50 languages, with pricing starting at $0.0035 per minute.
- Meta and Google are competing to release the first widely used AI digital assistant, each leveraging distinct strengths.
- New York City proposed rules requiring businesses to disclose AI customer service interactions, with fines starting at $525.
- Fewer than 20% of AI-influenced decisions have clear ownership or accountability, according to a Sopra Steria Next report.
- Espressif launched the ESP-Mosaico development platform for edge AI applications, featuring WiFi 6, Bluetooth 5.4, and Zigbee or Thread support.
- Central bankers are increasingly discussing AI's impact on jobs and wages, with Asian economies like China and Indonesia facing risk from slowing AI demand.
- TAITRONICS 2026 in Taipei connects Taiwan's US$192 billion AI hardware supply chain with nearshoring hubs like Mexico.
- AI app creation is raising copyright questions, with the author expecting legal pressure and industry adaptation similar to past tech battles.
China's Moonshot AI expands operations into Singapore
Chinese AI startup Moonshot AI is expanding into Singapore with a new sales job listing on LinkedIn. The role seeks people with sales, consulting, or engineering experience to build revenue and partnerships. Moonshot already has a business entity in Singapore from July 2023. The company recently released its Kimi K3 model and is valued at $35 billion.
Moonshot AI posts Singapore sales role to reach global customers
Moonshot AI, valued at $35 billion after a July fundraise, posted a sales role in Singapore dated October 20. The company is also dismantling its offshore Cayman Islands structure to prepare for a possible Hong Kong IPO. Backers include Alibaba, Tencent, Meituan, and China Mobile. Moonshot's Kimi K3 is a 2.8-trillion-parameter model compared with frontier releases from OpenAI and Anthropic.
Espressif launches ESP-Mosaico modular platform for AI hardware
Espressif Systems released the ESP-Mosaico, an ESP32-S31 development platform with a 480 by 480 OLED touchscreen, haptic feedback, built-in battery, and USB Type-C. It targets edge voice, multimedia, and motion sensing applications. The platform includes a Camera Module with a 3.2 MP OV3640 sensor and a Multi-function Interaction Module with sensors and RGB LEDs. It supports WiFi 6, Bluetooth 5.4, and Zigbee or Thread connectivity.
Letter urges delay in advanced AI deployment until safeguards exist
A letter to the editor calls for slowing advanced AI expansion until regulations and safety frameworks are in place. The writer warns about risks to jobs, privacy, and democratic processes from deepfakes. The letter also highlights heavy energy and water use by data centers as an environmental concern. It urges policymakers and scientists to build guardrails before large-scale AI growth continues.
TAITRONICS 2026 opens in Taipei to connect global AI hardware chains
TAITRONICS 2026 opened in Taipei on October 25 as a major international trade show. It links Taiwan's US$192 billion AI hardware supply chain with nearshoring hubs like Mexico. The event promotes collaboration and innovation among industry players. Attendees can explore AI hardware, software, and solutions.
Meta and Google prepare for consumer AI assistant showdown
Meta and Google are competing to release the first widely used AI digital assistant. Google is seen as well positioned due to its AI models and wide product reach. Meta has an advantage because its social networks already reach billions of people. Both companies face the risk that AI agents could reduce advertising revenue by removing apps and websites from the user experience. Investors will watch closely for any slowdown in search growth linked to AI use.
New York City proposes rules against AI customer service loops
New York City has proposed rules requiring businesses to tell customers when they are talking to AI instead of a human. Companies must also offer a simple way to reach a real person. The rules would ban AI systems that give wrong or off-topic answers while promising customer service. Fines would start at $525 and rise for repeat offenses. The rules would apply to all businesses in the city, with some exemptions for airlines and national banks.
Most AI decisions lack clear ownership and accountability
A new Sopra Steria Next report says fewer than 20% of AI-influenced decisions have an identified owner or clear accountability. Fewer than 25% of AI decisions in production can be traced and replayed from start to finish. The report calls for a five-layer framework covering data, models, platforms, APIs and decisions. It urges leaders to define who controls and is responsible for automated choices. The report warns that unaccountable AI systems cannot be properly governed or scaled.
AI app creation raises copyright questions for tech firms
People are using AI to create apps based on copyrighted ideas, raising legal questions. The article compares this to past battles like browser wars, ad blockers and Napster. It suggests that blocking all AI-made apps would be very difficult. The author expects a mix of legal pressure, agreements between companies and user-created tools. Tech firms may eventually accept AI-created software, similar to how they adapted to ad blockers.
Mistral AI launches Voxtral 3 speech models and voice agent SDK
Mistral AI released Voxtral 3, a family of speech models and a voice agent SDK. The release includes three transcription models, one text-to-speech model and an SDK built on LiveKit. The models support diarization, context biasing and 50 languages. Pricing starts at $0.0035 per minute for batch transcription and $13 per million characters for speech output. The models are available on Mistral API and can run on user infrastructure.
Central bankers now focus on AI impact on jobs and wages
Central bankers are increasingly discussing how artificial intelligence affects jobs and wages. Donald Kohn, a former Federal Reserve vice chairman, said workers feel insecure about their jobs because of AI. Many Asian economies like China and Indonesia face risk if AI demand slows down. The IMF and World Bank meetings showed surprise at how strong the world economy remains. Central bankers plan to discuss AI's effects at upcoming meetings in Basel, Switzerland.
Sources
- China’s Moonshot AI to operate in Singapore, jobs post shows
- Moonshot AI Posts a Singapore Sales Job as It Chases Global Customers
- ESP-Mosaico – An ESP32-S31-based modular development platform for AI-assisted hardware creation
- Letter to the Editor: Delay The Deployment Of Artificial Int...
- Taiwan’s TAITRONICS 2026 Seeks to Advance Global AI Chains
- Meta and Google are about to have a consumer AI showdown
- New York City is proposing rules to ban AI customer service 'doom loops'
- AI: fewer than 20% of automated decisions have an identified owner
- People are vibe coding copyrighted apps with AI – will tech firms sue?
- Mistral AI Releases Voxtral 3: Speech-to-Text, TTS and a Voice Agent SDK
- AI and jobs: The hot new topic for central bankers
Comments
Please log in to post a comment.