China Just Rolls Out Qwen3👑

In just the past week, three tech giants have made bold moves that signal a new chapter in the age of AI—each reshaping how we interact, learn, and build with intelligent systems.

It starts with Qwen3, China’s latest open-source language model, making waves for its ability to switch between deep thinking and fast answers. Built to support 119 languages and work across platforms like Hugging Face and Ollama, it's not just another AI—it's a toolkit for researchers and developers around the world who want power and flexibility in one place.

Meanwhile, Meta is going personal with its new Meta AI app. Think of it as your own voice assistant that remembers what you like, responds in natural conversation, and follows you across devices—from your phone to smart glasses. Whether you're chatting, asking questions, or browsing its Discover feed, this assistant is designed to adapt to your life while giving you full control.

And then there’s Duolingo, taking a leap into the future of education by becoming an AI-first company. They’re not waiting for perfect tools—they're moving fast to build them. From scaling content to reimagining how people learn with features like AI-powered video calls, Duolingo is betting big on AI to bring teaching quality to new heights.

Together, these stories show a clear direction: AI isn’t just a feature anymore—it’s becoming the foundation

Qwen3 Launches as China’s Next-Gen Open AI Model

Qwen3 is the latest open-source large language model from the Qwen family, offering powerful AI with flexible "thinking" and "non-thinking" modes for deep reasoning or instant answers. With support for 119 languages, improved coding and agentic capabilities, and efficient performance even in smaller models, Qwen3 is designed for both developers and researchers. It’s easy to use on popular platforms like Hugging Face and Ollama, and can be fine-tuned for specific needs. Qwen3’s hybrid approach, extensive multilingual support, and open availability make it a versatile tool for building innovative AI solutions worldwide.

Meta AI App: Your Personalized Voice Assistant Across Devices

Meta has introduced the Meta AI app, a new personal assistant powered by the advanced Llama 4 model. Designed to learn your preferences and remember context, the app offers natural voice and text conversations, integrates with Facebook and Instagram for personalized responses, and connects seamlessly with Ray-Ban Meta smart glasses. It features a Discover feed where users can share and explore AI prompts, with full control over what they share. Voice interactions, including a full-duplex speech demo for more natural conversations, are initially available in the US, Canada, Australia, and New Zealand. The app syncs conversations across devices and the web, making Meta AI accessible wherever you are, while prioritizing privacy and user control.

Duolingo Launches AI-First Strategy to Transform Learning

Duolingo is officially becoming an AI-first company, embracing AI as the next major platform shift just as it did with mobile in 2012. The company recognizes that AI is transforming how work gets done and is committed to moving quickly rather than waiting for perfect technology. AI will help scale content creation, improve features like Video Call, and bring teaching quality closer to that of the best human tutors. This shift means rethinking workflows and building new systems designed for AI, with a focus on urgency and innovation. Duolingo will gradually reduce contractor work that AI can handle, incorporate AI use into hiring and performance reviews, and support employees with training, mentorship, and new tools. Despite these changes, Duolingo remains deeply committed to its employees, aiming to remove bottlenecks so they can focus on creative and meaningful work while staying ahead in AI-driven education.

Hand Picked Video

In this video, we'll guide you through running Qwen 2.5 locally and privately on your laptop.

Top AI Products from this week

• Daytona - Daytona Cloud reimagines infrastructure for AI agents with sub-90ms startup times, bare metal performance, and stateful execution—capabilities traditional clouds can't match. Create, control, and deploy AI agents with unprecedented speed and flexibility.

• Bruno - The API-client built for developers. Open-source. Local . Native git. Bruno stores collections on your filesystem which allows you to natively interact with git for collaboration and versioning.

• Exa Websets - Find perfect lists of leads, candidates or anything in the web. Websets’ semantic search lets you ask in plain English, it agentically sources results and enriches them with you ask for - like emails, tags, and more - all in the time it takes to brew a coffee.

• Transkriptor 2.0 - Transkriptor is your AI-powered note-taking and transcription tool — now smarter and faster with its latest version. Capture conversations, generate summaries, and extract insights effortlessly.

• Raindrop - AI engineers use Raindrop to get alerts about hidden issues and successes in their AI products. Raindrop sends alerts when your AI misbehaves and links straight to events, so you can dig into conversations or traces, understand the root cause, and fix it, fast

• Breaktime Buddy - Smarter breaks = better output. AI-powered recovery, micro-therapy, and focus resets in 5 minutes. Because doomscrolling isn't rest. Built for remote workers, creators, and tired brains everywhere.

This week in AI

• Tavus Hummingbird-0 - Tavus has launched Hummingbird-0, a state-of-the-art zero-shot lipsync model now in research preview. It delivers the most accurate, photorealistic lipsync results-open or closed source-requiring just one video and any audio, no training needed.

• Hugging Face SO-101 Robot Arm - Hugging Face’s SO-101 is a 3D-printable, programmable robotic arm with improved motors and camera, starting at $100; prices vary by supplier and assembly.

• Virtual Fly AI - AI-powered digital fruit fly replicates real walking, flying, and vision-guided behavior-advancing neuroscience with open-source simulation tools.

• Mobility AI Advances - Google Research’s Mobility AI uses AI-driven measurement, simulation, and optimization to improve urban transport, tackling congestion, safety, and emissions with data-driven solutions.

• Hana Interactive Design - Hana is a real-time, GPU-accelerated 2D design canvas enabling fast vector editing, state-based animation, and live collaboration-all production-ready and browser-native.