China's DeepSeek has disrupted the global market by releasing R2, an open-source model that rivals the reasoning capabilities of GPT-4o. With its efficient training process and massive cost reductions, R2 is empowering every developer to run state-of-the-art LLMs locally. This is a pivotal moment for the democratization of open-source intelligence.