Google's Robots Can Now Think, Search the Web and Teach Themselves New Tricks

Google DeepMind introduced two advanced AI models, Gemini Robotics 1.5 and Gemini Robotics-ER 1.5, enhancing the intelligence of robots. These models enable machines to reason, plan, and search the internet for information, shifting from mere command execution to adaptable problem-solving. With capabilities such as sorting laundry by color, packing suitcases based on weather forecasts, and checking local recycling rules, these robots can generalize knowledge across different tasks. The Gemini Robotics-ER 1.5 acts as the brain, creating plans and utilizing Google Search, while the Gemini Robotics 1.5 executes these plans in the physical world. The initiative demonstrates significant progress towards achieving general-purpose intelligence in robotics, allowing for a higher degree of autonomy and adaptability compared to traditional models that rely on explicit programming. The robots achieved a success rate of 20% to 40% in handling sorting tasks, showcasing potential for future improvements. This technology positions Google at the forefront of the robotics revolution, focusing on developing robots that are responsive to dynamic environments and capable of learning from experience.

Source 🔗