Qwen3 Just Landed

Welcome back apprentices! 👋

China just hit the AI gas pedal — and left the brakes back in Palo Alto.

While OpenAI and Google whisper about updates, Alibaba just cannonballed into the pool with six brand-new models — some smarter, some safer, all loud. 

No, you don’t need to memorize acronyms or benchmark scores. 

Just know this: the global AI race just got weirder, faster, and way more interesting.

In today's email

• Alibaba drops six new AI models — all at once

• Qwen3 beats top vision and math benchmarks

• China’s AI strategy: fast, open, and aggressive

• What this means for you

• + New AI Tool

_{Read Time: 4 minutes}

Quick News 

🛠️ Apple Talent, New Mission. OpenAI is raiding Apple’s hardware team and teaming up with iPhone manufacturers to build its first line of AI devices.Led by former Apple exec Tang Tan (with design guidance from Jony Ive), they’re luring top talent with $1M+ offers and promising fewer meetings, more moonshots. Products in the works? A speaker-like gadget, smart glasses, a wearable pin — all aiming for a 2026/27 debut.

💸 $100 Billion, 10 Gigawatts, and Infinite AI Ambition. Nvidia and OpenAI just announced the biggest AI infrastructure project in history — and it’s powered by more GPUs than most nations can imagine. Nvidia plans to invest up to $100 billion into OpenAI over time, deploying 10 gigawatts of systems (millions of GPUs!) to fuel the next generation of intelligence. The first wave goes live in 2026, and while Microsoft and Oracle are still in the mix, Nvidia is now OpenAI’s preferred compute partner.

🧑‍💻 Productive but Paranoid. According to Google Cloud’s new DORA report, 90% of developers now use AI tools daily — but a third of them barely trust the output. Even with skepticism, most still report better code and faster workflows, proving AI is less a silver bullet and more of a skeptical sidekick. Google also dropped a new “AI Capabilities Model” to help teams get smarter about using AI without handing over the keys.

Together with Fellow

Don’t be the one putting your company at risk 😳

Most meetings include confidential data - from project details to client information.

But not every AI notetaker protects that information with the security your company deserves.

If the AI notetaker you’re using trains their AI models on your team’s conversations, you could be putting your company at risk without realizing it.

This AI Meeting Notetaker Security Checklist helps you avoid that.

In just two minutes, you’ll learn the 7 checks to ensure your team’s AI meeting notes stay private and secure.

Don’t let your meetings become someone else’s dataset.

Get the checklist

Alibaba 
Alibaba’s AI Playbook: All Gas, No Brakes

While Western AI giants are playing it slow and surgical, Alibaba just unleashed a six-model blitz across text, vision, audio, and safety — led by the newly announced Qwen-Max, a 1T parameter model brushing up against frontier capabilities. 

The release includes multi-modal models, math-dominant variants, real-time speech translation, and new safety frameworks — all in one go.

At a time when OpenAI, Anthropic, and Google drip-feed updates, Alibaba is carpet-bombing the model landscape, building momentum and buzz across Asia and beyond. With Qwen3, China may finally have an answer to the GPT-4/Gemini Claude-class — and it's launching fast and loud.

Key Points:

• 🧠 Qwen-Max (1T model) shows near-frontier strength in agentic tasks and coding; Heavy variant aces math reasoning benchmarks (100% on GSM8K, SVAMP).
  
  

• 🎥 Qwen-Omni handles text, image, audio, and video, with speech recognition in 19 languages and generation in 10 — all in a single model.
  
  

• 👀 Qwen-VL leads among open-source vision models and beats closed models like GPT-4V in visual benchmarks (e.g. SEED-Bench, MME).
  
  

• 🗣️ LiveTranslate-Flash offers near-instant real-time interpretation, built for calls, meetings, and cross-lingual interactions.
  
  

• 🛡️ Guardmodels are designed for safety and moderation, covering prompts, output filtering, and model-level alignment.
  
  

• 🧑‍💻 Qwen-Coder V2 upgraded to better support LLM dev workflows, outperforming open competitors on code synthesis and reasoning tasks.

Alibaba vs. the Western Giants

➡️ Alibaba’s strategy is more “AI app store,” while OpenAI/Google lean into the all-in-one superbot approach.

Follow the Money

After unveiling the Qwen3 stack, Alibaba’s U.S.-listed shares jumped 6% in regular trading, with an additional 9% pop in premarket after CEO Eddie Wu pledged "sizable AI spending increases". 

This isn't just an R&D splash — it's a strategic pivot, signaling confidence in becoming a global AI contender.

Meanwhile, investments in LLM research from China's top firms (Alibaba, Baidu, Tencent) continue to accelerate — often with state-aligned incentives that reduce regulatory friction vs. U.S. firms.

The Bigger Picture

Just months after DeepSeek’s R1 shook the open-source math benchmarks, Alibaba’s Qwen3 drop positions it as China’s closest challenge yet to Western frontier labs. The models aren’t just big — they’re increasingly specialized, accurate, and built for commercial use.

This all plays into a wider macro-shift:

• The East is scaling AI faster, with more aggressive public-private investment.
  
  

• The West is slowing down — in part due to regulatory caution, safety debates, and monetization challenges.

If Qwen’s approach gains traction globally, it could reshape the open vs. closed-source AI race — and offer developers more flavors to build with.

So… What’s the Deal for You?

Whatever you do, this matters a lot. More models mean:

• More specialized tools (don’t force GPT-4 to do everything)
  

• Potentially lower costs (open models = fewer API fees)
  

• Regional balance in model performance (language support, local reasoning)

It also means Western AI dominance is no longer guaranteed — and the smartest builders will keep tabs on both sides of the globe.

Don’t wait for OpenAI to drop a new tool — go poke around the Qwen open models on Hugging Face or test their math/coding demos. Think of it like trying a new IDE or framework: it may be rough, but could give you a speed edge no one else is using yet.

Help Your Friends Level Up! 🔥

Hey, you didn’t get all this info for nothing — share it! If you know someone who’s diving into AI, help them stay in the loop with this week’s updates.

Sharing is a win-win! Send this to a friend who’s all about tech, and you’ll win a little surprise 👀

Today’s Toolbox

Ever heard of Agentic Commerce?

Every Shopify store will face Agentic Commerce eventually. The question is: are you ready?

Zipchat.ai turns traffic into profit with instant answers, cart recovery, and 24/7 support automation. Add the AI Chat-Bubble to your store or deploy it on WhatsApp to win Q4.

Use code NEWSLETTER10 for 10% off forever.

Try Zipchat free

🧪 Test the Prompt

A playground for your imagination (and low-key prompt skills).

Each send, we give you a customizable DALL·E prompt inspired by a real-world use case — something that could help you in your business or job if you wanted to use it that way. But it’s also just a fun creative experiment.

You tweak it, run it, and send us your favorite. We pick one winner to feature in the next issue.

Bonus: you’re secretly getting better at prompt design. 🤫

👑 The winner is…

Last week, we challenged you to test GPT-4o’s visual generation skills with this prompt.

Here’s the WINNER:

Congrats to Louis for his creation!🥳

Want to be featured next? Keep those generations coming!

🎨 Prompt: The Autonomous Expo Pavilion

Design a next-generation expo pavilion built to showcase the future of [specific industry or field]. The architecture is sleek and futuristic, using clean geometric forms and advanced materials. The main structure features kinetic panels that react to sunlight, and autonomous robots guide visitors through interactive installations. A large translucent display shows real-time simulations related to the industry.
Style: ultra-crisp architectural rendering, hyper-realistic materials, dynamic lighting, grounded sci-fi realism, minimal but futuristic.

We’ll be featuring the best generations in our next edition!

_FEEDBACK