Quantum Physicists 'De-Censor' DeepSeek R1; Google Unveils Gemini 3 with Agent Capabilities

Quantum physicists at Multiverse Computing, a Spanish firm, have reportedly stripped the censorship from DeepSeek R1, a powerful reasoning AI model originally developed in China. Due to regulations, Chinese AI companies often incorporate censorship layers to align content with laws and socialist values. Multiverse Computing used quantum-inspired AI techniques to create DeepSeek R1 Slim, which is 55% smaller than the original but maintains comparable performance. This allowed them to remove Chinese censorship, enabling the model to answer sensitive questions like Western models. Google has unveiled Gemini 3, a significant upgrade to its flagship multimodal model. The company claims the new model exhibits enhanced reasoning capabilities and more fluid multimodal functionality, enabling it to operate across voice, text, and images. Gemini 3 will also function as an agent, with an experimental feature called Gemini Agent designed to handle multi-step tasks directly within the app. It can connect to services like Google Calendar, Gmail, and Reminders, allowing it to manage schedules and organize inboxes upon user authorization.