Nvidia has announced GR00T (Generalist Robot 00 Technology), a new artificial intelligence model designed to give robots the ability to perform tasks without having to pre-program each behavior

This new model is part of a "moonshot mission" by Nvidia's GEAR Lab to solve the problem of embodied AGI in the physical world This is essentially to give future superintelligence a leg up and ensure that it is not confined to a black box

GR00T was unveiled at GDC by Nvidia CEO Jensen Huang, who brought a pair of cute Disney robots on stage that learn to walk themselves using Nvidia's AI

Huang said that the GR00T is part of a new industrial revolution and that being versatile will allow the robots to learn tasks by watching humans In turn, this means that they will be able to repeat those tasks on assembly lines and in spaces that are dangerous for humans

GR00T and other projects like it are part of the solution to the Embodied AI problem This is a growing area of research that aims to give foundational AI models like ChatGPT and Google Gemini the ability to interact with the real world

Part of this comes in the form of virtual environments Google's recent SIMA game agents are one example, but it also includes bringing them into our world

We are already seeing companies developing a variety of robots in this area They do jobs that are either too dangerous or too tedious for humans

Unicorn startup Figure has partnered with OpenAI to give its Figure 01 humanoid robot the ability to reason and learn while interacting with humans

Google's Deep Mind is working to train robots to think for themselves, giving them the ability to decide how to move, act, and complete tasks

Nvidia's plan is to create a foundational framework that can be used by a wide range of companies producing humanoid robots

The GR00T in this scenario is an AI framework, not a living tree traveling through space with a talking raccoon; built by GEAR Labs, the GR00T is built on Nvidia's deep technology stack, training, scaling, and includes other models for powering up AI models on the move

Jim Huang, Nvidia's GEAR group leader, wrote in X that the GR00T "enables the robot to understand multimodal instructions such as language, video, and demos, and perform a variety of useful tasks"

He stated that they are working with various leading humanoid robotics companies to enable GR00T to work with different body types and ecosystems

Fan and Yook Zhou, co-leader of GEAR, wrote in X:" GR00T enables robots to follow natural language instructions and learn new skills from human videos and demonstrations"

The goal is to create a generalist robot with both a versatile body and an intelligent mind This will allow millions of these robots to be deployed for real-world tasks in the future, eliminating the need to program and specially train each machine for each task