← All Issues
Thursday, June 11, 2026

AI Daily

Your Automated Intelligence Briefing

Frontier Models

The 'Too Dangerous' Model Is Now Public. Keep an Eye on the Leash.

Anthropic has released Claude Fable 5, a Mythos-class model, to every paying customer, two months after insisting that capability was for vetted cyberdefenders only. The release comes with safety classifiers that hand risky questions to a weaker model, a price double that of Opus 4.8, and a deadline: it leaves subscription plans on June 23.

Read full story →
Economy

The AI Boom Is Now Holding Up China's Economy. That Cuts Both Ways.

China's May exports jumped 19.4 percent, smashing forecasts, with chip shipments up 111 percent and computers up 66 percent. But furniture, toys and footwear are flat or falling, retail sales are near zero, and economists warn the AI windfall is propping up a K-shaped economy while delaying the rebalancing Beijing keeps promising.

Read full story →
Geopolitics

Beijing's $295 Billion Plan to Compute Without Nvidia

China is drafting a five-year, 2 trillion yuan plan for a national AI data centre grid running on at least 80 percent domestic technology, locking out Nvidia and AMD. The money is the easy part: SMIC's fabs are already above 93 percent utilisation, and even Chinese executives doubt local silicon can fill the racks by 2028.

Read full story →