5 things to know about Nvidia GR00T, the humanoid robot AI model

Nvidia earlier this week launched a general generative artificial intelligence (AI) platform for humanoid robots called GR00T. The platform features hardware and software solutions for humanoid robots that equip them with capabilities that enable them to perform complex tasks and learn new tasks with human-assisted input. The AI ​​model is paired with the company’s new chipset called Jetson Thor, which will power the AI ​​features. Nvidia revealed that the GR00T was developed with the help of its Isaac robotics platform, which has also undergone multiple upgrades to enable end-to-end development of humanoid robotics platforms.

Announce GR00T is the abbreviation of Generalist Robot 00 Technology. Nvidia founder and CEO Jensen Huang said, “Building basic models for general humanoid robots is one of the most exciting problems in the field of artificial intelligence today.” This statement carries decades of research on The result of research into humanoid robots that not only look like us, but also function like us.Although researchers have built robots that are very close in appearance and some functions, such as playing football, robot Developed at the University of California, Los Angeles (UCLA), ARTEMIS’s goal has always been to develop a software platform powerful enough to handle complex human movements and gain knowledge about the world around it. With the GR00T, Nvidia seems to be taking a big step in this direction. Here are five things you should know about this landmark development.

Nvidia founder and CEO Jensen Huang launches GR00T
Image source: YouTube/Nvidia

GR00T artificial intelligence model

It would be easy to mistake the GR00T for a humanoid robot, considering that multiple robots from Disney projects were on stage with Huang during his Nvidia keynote at the GPU Technology Conference (GTC) in San Jose. However, GR00T is not a robot. Instead, it can be understood as the brain of a robot.

What Nvidia has created is a generative AI model that can power bots, just like OpenAI’s GPT-4 powers ChatGPT and Google Gemini Pro powers the Gemini chatbot. Artificial intelligence models contain algorithmic architectures and data sets that enable them to generate information and perform specific tasks. This AI model became the basis for the chatbots and AI tools we ended up using, also known as the base model. Therefore, the GR00T is the base model.

What does GR00T do?

As the brain of the robot, Nvidia’s GR00T AI model is responsible for receiving input and processing output. Nvidia says the base model can take text, speech, video and live human demonstrations as input. This means that any GR00T-powered robot can receive verbal instructions to do something, be guided through written text, video, or even be shown how to do something.

During the demonstration, Huang told a robot to move toward him, which it did. Ideally, a robot could watch a human do push-ups and perform them itself. Video input will work the same way, and watching a drummer improvise should also teach it how to play the drums, but may not be as efficient.

The role of Jetson Thor chipset

While GR00T is the most important part of the humanoid robot software platform, it is not the only part. At the event, Jensen also launched a chipset called Jetson Thor. This chipset will power the robot. The AI ​​SoC is based on the Nvidia Thor chipset and is built on the company’s Isaac robotics platform. The company says the chip is powerful enough to help the GR00T run seamlessly. The SoC comes with specific enhancements for generative AI models as well as simulation and AI workflow infrastructure tools.

Isaac Robot Platform

Another important component of the GR00T ecosystem is the Isaac robotics platform. The platform has also been upgraded with tools that enable end-to-end development of AI-based bots. The process includes development, simulation and deployment, using specific digital platforms and computers to power it. Let’s break it down.

Huang stood next to one of the robots developed using GR00T
Image source: YouTube/Nvidia

While the GR00T AI model enables humanoid robots to learn on the job, the robots should have some understanding of the world in which they will perform these tasks and a basic understanding of how to perform them. To do this, a training ground for robots needs to be created where they can learn how to become robots before going out into the world.

That’s what the Isaac robotics platform does. This is a gym for artificial intelligence robots that deploys functional robots that are ready to run through virtual world simulations and real-world training. For GPU-driven simulations, Nvidia uses its new Isaac Lab, which is based on Isaac Sim platform.Additionally, Nvidia is using its OSMO computing platform and OVX and DGX systems for training robots simultaneously.

The future of robotics

Nvidia’s GR00T AI system is a major leap forward in robotics technology for training and deploying something close to so-called artificial general intelligence (AGI). For those unfamiliar, AGI is a branch of artificial intelligence that develops capabilities that are equivalent to those of humans. These include cognitive tasks and physical tasks.

OpenAI has also recently begun work on AGI in partnership with 1X Technologies and Figure, two robotics startups that will use the AI ​​company’s models to power their machines.

However, it’s still early days for humanoid robots. Despite warnings of doom from artificial intelligence researchers that we may end up in a future world where the artificial intelligence Skynet becomes human overlord in the “Terminator” movies, there’s still a long way to go. Nvidia and OpenAI are both in the early stages of developing robotics technology, and currently, the robots they power are unlikely to be able to handle complex tasks without human supervision. But this is just the beginning, in ten years we may see humanoid robots entering the workforce and working alongside humans. No one could predict how this would change the world.