🇪🇺 EU launches $1.1B plan to reduce AI dependence on US and China 🇪🇺
AI strategies target sovereignty and research capabilities

The European Union unveiled a $1.1 billion plan featuring twin strategies to ramp up AI adoption and research across the bloc. The "Apply AI Strategy" aims to reduce dependency on US and Chinese technologies in critical economic sectors, while the parallel "AI in Science Strategy" boosts research capabilities through a virtual resource hub. The comprehensive approach addresses both immediate commercial needs and long-term research competitiveness as Europe seeks greater AI autonomy.

🎬 Sora downloads nearly match ChatGPT's launch 🎬
OpenAI's video generator achieves 627,000 downloads despite invite-only access

Sora reached 627,000 downloads in its first week compared to ChatGPT's 606,000, despite being invitation-only and requiring a paid subscription. The video generation app hit No. 1 on the US App Store by day three, demonstrating massive consumer appetite for AI video tools. This performance signals video generation could match text AI's mainstream adoption trajectory, with Sora proving that premium AI tools can still achieve viral-level downloads when the technology delivers transformative capabilities.

💼 IBM integrates Claude across applications 💼
Tech giant expands AI capabilities with responsible deployment focus

IBM is adding Claude to its applications, starting with developer tools, and partnering with Anthropic to guide responsible AI deployment across the enterprise. The collaboration builds on Anthropic's growing momentum in enterprise deals following major partnerships announced this week. The integration represents IBM's commitment to expanding AI capabilities while maintaining focus on ethical and responsible artificial intelligence implementation in business environments.

INTERESTING TO KNOW

🧠 Samsung's 7M model beats OpenAI's o3-mini - on specific problems 🧠

Samsung researcher Alexia Jolicoeur-Martineau developed the Tiny Recursion Model (TRM) with just 7 million parameters that reportedly outperforms much larger models like OpenAI's o3-mini on complex reasoning tasks. Rather than relying on raw computational power, TRM repeatedly refines its own predictions, challenging the industry assumption that larger models automatically deliver better performance.

🇨🇳 Tencent claims No.3 global spot 🇨🇳

Tencent released Hunyuan-Vision-1.5-Thinking, a new multimodal vision-language model that secured the No. 3 position on LM Arena's competitive Vision Arena leaderboard. The Chinese tech giant's achievement demonstrates continued advancement in AI capabilities despite ongoing geopolitical tensions and trade restrictions.