OpenAI strikes a rushed deal with the Pentagon to deploy AI in classified military settings while claiming safeguards against autonomous weapons and mass surveillance, raising urgent questions about whether safety measures can be maintained amid geopolitical tensions and military urgency. Simultaneously, Google releases a wave of new models including Gemini 3.1 Flash-Lite, and AI systems achieve a breakthrough in solving PhD-level math problems faster than humans can formulate the tests, marking a significant inflection point in AI capabilities. These converging developments—military deployment, rapid model proliferation, and superhuman performance on complex reasoning tasks—collide at a moment when the AI safety community remains divided over whether companies are prioritizing speed and competitive advantage over the safeguards needed for responsible deployment.