Google DeepMind has introduced a new AI model called Gemini Robotics On-Device, which enables robots to perform tasks locally without needing an internet connection.
Features of Gemini Robotics On-Device
The Gemini Robotics On-Device model builds on the previous Gemini Robotics AI model and can control robot movements. It is designed to perform tasks directly on devices, providing greater autonomy. Developers can control and fine-tune the model using natural language prompts.
Comparison with Previous Models
According to Carolina Parada, head of robotics at Google DeepMind, the new model provides offline features nearly as well as the cloud-based flagship. While the hybrid Gemini Robotics model remains more powerful, this on-device model shows impressive performance for applications with limited connectivity.
Competition in Robotics
Other companies are also expressing interest in robotics. Nvidia is developing a platform for creating foundational models for humanoids. Hugging Face is working on open models and datasets while advancing its own robotics developments. The interest in robotics is growing and raises questions about the future of the technology.
The launch of the Gemini Robotics On-Device model represents a significant advancement in autonomous robotic systems capable of efficiently operating in low-internet conditions while highlighting the increasing competition in the AI and robotics field.
