Nvidia Says 'the Age of Generalist Robotics Is Here' (theverge.com) 10
During the company's GTC 2025 keynote today, Nvidia founder and CEO Jensen Huang announced Isaac GR00T N1 -- the company's first open-source, pre-trained yet customizable foundation model designed to accelerate the development and capabilities of humanoid robots. "The age of generalist robotics is here," said Huang. "With Nvidia Isaac GR00T N1 and new data-generation and robot-learning frameworks, robotics developers everywhere will open the next frontier in the age of AI." The Verge reports: Huang demonstrated 1X's NEO Gamma humanoid robot performing autonomous tidying jobs using a post-trained policy built on the GR00T N1 model. [...] Other companies developing humanoid robots who have had early access to the GR00T N1 model include Boston Dynamics, the creators of Atlas; Agility Robotics; Mentee Robotics; and Neura Robotics. Originally announced as Project GR00T a year ago, the GR00T N1 foundation model utilizes a dual-system architecture inspired by human cognition.
System 1, as Nvidia calls it, is described as a "fast-thinking action model" that behaves similarly to human reflexes and intuition. It was trained on data collected through human demonstrations and synthetic data generated by Nvidia's Omniverse platform. System 2, which is powered by a vision language model, is a "slow-thinking model" that "reasons about its environment and the instructions it has received to plan actions." Those plans are passed along to System 1, which translates them into "precise, continuous robot movements" that include grasping, moving objects with one or two arms, as well as more complex multistep tasks that involve combinations of basic skills.
While the GR00T N1 foundation model is pretrained with generalized humanoid reasoning and skills, developers can customize its behavior and capabilities for specific needs by post-training it with data gathered from human demonstrations or simulations. Nvidia has made GR00T N1 training data and task evaluation scenarios available for download through Hugging Face and GitHub.
System 1, as Nvidia calls it, is described as a "fast-thinking action model" that behaves similarly to human reflexes and intuition. It was trained on data collected through human demonstrations and synthetic data generated by Nvidia's Omniverse platform. System 2, which is powered by a vision language model, is a "slow-thinking model" that "reasons about its environment and the instructions it has received to plan actions." Those plans are passed along to System 1, which translates them into "precise, continuous robot movements" that include grasping, moving objects with one or two arms, as well as more complex multistep tasks that involve combinations of basic skills.
While the GR00T N1 foundation model is pretrained with generalized humanoid reasoning and skills, developers can customize its behavior and capabilities for specific needs by post-training it with data gathered from human demonstrations or simulations. Nvidia has made GR00T N1 training data and task evaluation scenarios available for download through Hugging Face and GitHub.
Sure, right after ray tracing finally arrives (Score:2)
Marketing hype is growing, the lead time is like ten years at least now, at these prices the current technology is like going back to the future. The only people buying into this are investors who don't really understand the business they are investing in. This salesmanship is just massively overselling the value of these classist transnational corporations. It can't be long before this house of financial cards collapses once again. Like every game of monopoly, anyone caught without real property will be ta
Re: (Score:2)
Ray tracing on GPUs are already quite good. Good enough for anything, anyway. I mean nowadays real-time path tracing seems to be what graphics nerds are on about. Robotics will keep developing for as long as people need work done and hate doing it. Farming needs robots, every factory job is a target for robotic takeover as well.
Re: (Score:2)
This is a bad answer to a bad analogy. Nvidia's cards don't do ray tracing. That's why they renamed ray tracing "path tracing" in their marketing. Stop parroting marketing terms.
But robotics development based on current models, unlike nvidia's ray tracing marketing wank isn't just marketing. It's actually actionable in real world in a very meaningful way, and has been for a while. For example until very recently Amazon burned billions of dollars on getting a robot to person a job of a picker. That is to pi
Re: (Score:2)
I couldn't disagree more. Yes, robotics is coming along nicely, so long as it's confined to a predictable environment. Look at the self-driving car tech, yah it's great, until it isn't. Then what? Are we going to hold the corporation that designed the robot accountable for its actions? Where is the accountability? We can already see corporations are not going to be responsible enough. The technology may improve but until the ethical issues are resolved, this is looks like it's going to be a dystopian future
Re: (Score:2)
Ray tracing on GPUs are already quite good. Good enough for anything, anyway.
Really, how many times do people turn off ray tracing so framerates are acceptable? See any competitive gaming with raytracing active? No? Me either. Just saying.
As for robotics, well we do have vacuum cleaners, sort of, they work ok until they run into crap
Re: (Score:2)
What do you mean "ray tracing finally arrives"? It objectively arrived several years ago. It's a think many games do. Just like AI training models for physical objects is a thing that objectively is already here. In fact a paper by NVIDIA's researchers postulated how to properly train such a model several years ago before the kids thought AI was cool.
Like every game of monopoly, anyone caught without real property will be taken off the board.
They have real property. They are selling proprietary AI models for robotics. I feel like you're on a general rant that has nothing at all to do with the subje
Re: (Score:2)
This general rant is the reason we don't have effective ray tracing or robotics yet and why we have marketing hype instead.
If the upper class had invested our capital into research and development instead of market manipulation and monopolization, we would be looking at a bright future instead of this dystopian nightmare. Just saying.
My advice to people is pay off your mortgage and get out of debt now, before it's too late once again.
Just because it's a rant, does not mean it's wrong,
Nvidia says a lot of things (Score:2)