🧋 Hello human and artificial friends,

Good news for us Germans — and our Indian friends: ChatGPT (soon to be ChatXXX) is about to save our love lives.

Our two languages have been ranked among the least sexy in the world.

But that’s about to change. LLMs are finally growing up and can soon generate erotic content.

👉🏻 Let’s put that to the test:
Prompt: “ChatGPT, how would a Frenchman flirt with a pretty mademoiselle at a café?”
Output (no joke): “Si tu continues à sourire comme ça, je vais devoir t’inviter à un verre.”

Looks like we still need to give it a little more... training.

P.S. Job posting of the week:  OpenAI is looking for a Solutions Engineer in Tokyo

Asia is where AI moves fastest. Every day, new tools drop. Most vanish overnight, but a few redefine how people live and work. We find those gems so you can watch, learn, and build your own.

Today’s idea

Asia is where AI moves fastest. Every day, new tools drop. Most vanish overnight, but a few redefine how people live and work. We find those gems so you can watch, learn, and build your own.

Today’s idea

AI Hay is a Vietnamese AI Q&A social app that combines generative AI with local sources and a real community. It has crossed 15 million downloads and handles over 100 million questions each month. 

In Sensor Tower’s State of AI Apps 2025 (Think “Forbes for the App Store and Google Play.”), it’s cited among the top AI apps and highlighted as the only Southeast Asian entrant in the top five cohort for H1 2025.

The company raised 10 million USD in July 2025 led by Argor Capital with Square Peg, Northstar Ventures, AppWorks, and Phoenix joining. 

What’s the USP?

AI Hay gives Vietnam-specific answers you can trust and verify.

✅ Faster, trusted answers for a whole country. People finish school and everyday tasks faster because answers are clear, localized, and come with sources.

✅ Beats global tools on local jobs. It understands Vietnam’s terms and rules, so users get the right method or steps the first time.

✅ Quality improves every week. Sources are visible and the community fixes mistakes. Good answers get better instead of going stale.

✅ A repeatable playbook. Focus on one market. Let the community maintain the canon. This combo wins habit without huge ad spend.

👓 AI glasses: the next big thing? Better put yours on!

🇨🇳 China’s frames of the future

China’s Alibaba readies Quark AI Glasses for launch. Runs Alibaba’s Qwen model and Quark assistant; deep hooks into Alipay pay-by-look, Amap navigation, Taobao price checks, Fliggy travel prompts. Real-time translation, hands-free calling, meeting transcription.

Rokid gains early traction in China. ~49 g frame with Micro-LED waveguide display, 12 MP first-person camera, integrated audio. Features real-time translation, object recognition, AR overlays, live transcription; built on Qualcomm AR1. Company cites 300k units sold in China. 

Baidu’s Xiaodu enters the fray. lightweight design with 16 MP ultra-wide camera + AI stabilization, 4-mic array, open-ear speakers. “Walk-and-Ask” on-the-go Q&A, object and calorie recognition, translation, smart memos; up to 56 h standby / 5 h audio.

🇺🇸US yardstick

Meta’s Ray-Ban Display. In-lens full-color microdisplay for private overlays; Neural Band wristband uses EMG to read tiny wrist/finger signals for hands-free control; 6-mic array with open-ear speakers and 12MP camera. It shows WhatsApp/Messenger/Instagram messages on-lens with live translation captions.

Watch

👀 Analysts see AI glasses as the most natural way to use a personal AI—always on, hands-free, and context-aware. The US–China race will be decided by model quality and ecosystem pull, not gadgets alone. Real hurdles remain—battery, heat, and input UX. Phones stay primary for now, but the category is shifting from flashy demos to real products on both sides.

🧠 China’s new robot looks you in the soul: In a Shanghai lab, startup AheadForm Technology has revealed Elf V1, a humanoid robot so lifelike it blinks, squints, and mirrors your expressions in real time. It has synthetic skin that wrinkles naturally and 30 tiny motors that move its face with eerie precision. Using cameras and microphones, it reads tone, gaze, and emotion to respond like a human conversation partner. Founder Hu Yuhang, a Columbia PhD, says the aim is not muscle power but emotional connection. The robot could one day assist in elder care, classrooms, or therapy rooms. Its realism is both fascinating and unsettling, showing how close machines are coming to feeling almost human.

🤖 Japan’s convenience stores now restock from Manila: In Tokyo’s FamilyMarts and Lawsons, shelf-stocking robots fill drinks and snacks — but when one drops a can, a worker 3,000 kilometers away in the Philippines picks it up using a VR headset. The robots, made by Tokyo startup Telexistence and powered by Nvidia and Microsoft systems, are remotely operated by Filipino “pilots” employed by Astro Robotics. Each supervises around 50 bots, stepping in when the AI stumbles. The setup lets Japan ease its labor shortage without expanding immigration and gives young Filipino engineers new tech jobs, though at local pay levels of around $300 a month. Their movements also train the AI to eventually work without them — raising the question of whether they’re building the future or coding themselves out of it.

📺 Samsung brings Perplexity's AI to the TV: Samsung has inked a deal with Perplexity to install its AI engine on its latest smart TVs. Users can choose between Perplexity, Microsoft's Copilot, and Samsung's own TV-focused AI when they press the AI button. This underlines that the shared living room experience is becoming a key battleground in the AI war, although it remains a challenging space for tech innovations beyond traditional media consumption.

🇨🇳 China: Robotics firm founded by Geely chief's son to shut down despite raising tens of millions of US dollars. More on that.

🇰🇷 South Korea: The government’s AI-powered textbook rollout has failed after widespread technical issues and public backlash. More on that.

🇭🇰 Hong Kong: A Cisco survey ranks the city lowest among 30 global markets for workplace AI readiness, with only 2% of firms classified as “pacesetters” in AI adoption. More on that.

This is the number of people in China now using generative AI — more than double the amount six months ago.

🇨🇳 AI goes mainstream: 36.5% of all Chinese internet users have adopted generative AI tools. Third-quarters of them are under the age of 40, with almost 37.5% having a degree of higher education.

🧩 Homegrown power: With OpenAI and Google DeepMind still blocked, domestic models like DeepSeek, Alibaba Cloud’s Qwen, and ByteDance’s Doubao dominate daily use and creative output. Beijing has attributed the surge in GenAI use to its nationwide “AI Plus” initiative, a state-backed program aimed at integrating artificial intelligence across industries and daily life.

🔎 Watch: The growth underscores China’s ability to scale technology fast under state direction, but also highlights how a closed ecosystem can accelerate adoption without necessarily ensuring openness, competition, or global interoperability.

🎨 即梦4.0 (Seedream 4.0) is ByteDance's new AI image creation model.
It combines text-to-image generation and advanced editing into a single system. Designers, marketers, and storytellers can use it to produce professional, large-scale visual content.
Its edge: Ultra-fast 4K generation powered by a highly efficient architecture, precise editing through language commands and native support for both Chinese and English prompts and text rendering.

📸 小云雀 (Xiao Yunque) is a multimodal AI assistant. It’s made for creating short-form content without needing any editing or design skills.
The app allows users to generate videos, design graphics, and complex image edits using simple commands.
Its edge: It automatically drafts video scripts, adds music, and performs color grading. It features Digital Human capabilities to create videos from text, image editing, and core features like intelligent video generation (15-60 seconds).

🎧 通义万相 (Tongyi Wanxiang) is Alibaba Cloud's AI Creative Generation Platform for instantly producing high-quality visual content. It automatically generates and edits visual content, allowing users to create stunning media and even realistic faces.
Its edge: Comprehensive Multimodal Generation, high-fidelity video generation (supporting up to 10-second, 1080P, 24fps videos), and it supports native audio-visual synchronization to generate video content with perfectly matched sound. 

🚗 Honda is doubling down on autonomous driving — with a little AI help from the US: The automaker has increased its investment in California-based startup Helm.ai, deepening a partnership that began in 2019. The two are co-developing “end-to-end” self-driving software that lets AI handle everything from perception to steering.

🔋 CATL invests in Chinese energy AI startup DaMao: The subsidiary of battery giant Contemporary Amperex Technology leads a nearly CNY100 million ($14 million) fundraising round for DaMao Technology, an AI startup in the energy sector. With the funding, DaMao aims to create 'virtual power plants' within data centers.

🤖 Shutu Tech wants robots to learn like humans — by watching videos: The Shenzhen startup just raised about 4 million USD in angel funding. Its system, SynaData, turns everyday internet clips into robot training data, pulling out details like hand movements, object paths, and 3D shapes. The result is faster, cheaper learning for robots.

✅ Read how one of Asia’s greatest visionaries envisions the future. *
✅ See what the new Xiaomi phone looks like. 👀
✅ Run as fast as you can if you see this AI police robot.
✅ Get a glimpse of how China is already living in the future — and read this travel guide first. Enjoy 15% off with code asiabitsfriends. **

_{*Yes, this is an affiliate link. No, you won’t pay a cent more. But it helps support asiabits. Thanks :)
** A message from our sponsor.}

🌏 Microsoft: Solution Engineer - Copilot, Hong Kong

🌏 NVIDIA: Product Manager - Open Models, Shanghai

🌏 IBM: Brand Partner Specialist - Automation, Seoul

🌏 OpenAI: Solutions Engineer, Tokyo

🌏 Sony: Intern - AI Application Engineer, Tokyo

🌏 ServiceNow: Software Engineer - Applied AI, Singapore

🌏 Shopee: Senior AI Engineer, Singapore

🌏 Grab: Intern - Software Development, Singapore

🌏 ATI JSC: AI Engineer, Hanoi

🌏 Manulife: Machine Learning Engineer, Chengdu

🐕 The good boy grillmaster: Forget fine dining, this entrepreneur is all about the Lamb Kebab! And at ¥1 per stick, you can't paw-ssibly say no.