The line between man and machine is becoming increasingly thin. Fourier, the Chinese robotics company, has just raised the curtain on its latest technological development: GR-II. This humanoid robot isn’t your typical incremental upgrade. It’s a leap forward that makes its predecessor, the GR-1, seem almost primitive by comparison. And to think, the GR-1 was already a technological marvel when it was introduced in 2023. But Fourier isn’t resting on its laurels. In just six months, it has transformed the GR-1 from a senior assistant to a full-fledged robotic all-rounder.
The new GR-II presents itself as a more refined and versatile product, ready to serve a wider and more varied clientele. No longer confined to the role of a hi-tech caregiver, GR-II aspires to become a universal assistant, capable of integrating into multiple contexts of our daily lives.
Machine Vision: The Eyes of the Future
One of the most surprising innovations of the GR-II is its artificial vision system. Fourier has implemented a system based exclusively on cameras, providing the robot with a 360-degree perception of the environment. It uses transformer models (I'll tell you what they are at the end of the paragraph, for those who don't know) to "translate" the environment around it into a set of spatial information, which it then reads and interprets to perfectly understand what is around it. This technology allows the Fourier GR-II to create a three-dimensional map of its environment in real time, identifying objects and people with almost human precision. It's as if it had developed a sixth sense, or rather: a "general sense".
Un Transformer model is a type of neural network architecture that has revolutionized the field of artificial intelligence, particularly natural language processing (NLP). First introduced in 2017 by Google Brain researchers in the famous paper “Attention Is All You Need,” Transformer is notable for its use of the attention mechanism, which allows the model to give more weight to specific parts of the input based on their relevance. In the case of Fourier, Transformer models have been adapted for computer vision tasks, such as image recognition and generating realistic images. If you want to learn more, there is some good content here.
Mobility and Dexterity: A Mechanical Dance
GR-II’s locomotion has been significantly improved. If GR-1 moved with the clumsiness of a newborn, GR-II moves with the clumsiness of a guy with diarrhea (next time I’ll find a better comparison, today I came up with this). It can perform complex movements such as half squats, waist twists and even air punches. “Brute strength” has not been neglected either: GR-II can lift up to 50 kg, a capacity that makes it not just an assistant, but a real collaborator in tasks that require physical effort.
However, the real magic of Fourier GR-II lies in its ability to interact with humans. The integration of voice interaction systems and brain-computer interfaces opens up previously unimaginable scenarios. GR-II does not simply execute commands, but can understand context, interpret emotions and respond appropriately. It is like having an assistant that not only understands what you say, but (in a rudimentary way) also how you feel.
Fourier and the co-created future
GR-II is not just a product, it is a technological manifesto. It represents Fourier's (and not only) vision of a near future in which, thanks to theembodiment and in imitative learning, humans and robots work together in harmony, each enhancing the other's capabilities. The question is no longer "if" it will happen, but "how" we will manage this new era of human-machine coexistence. GR-II and its "kind" are knocking on the door of our present. And they intend to stay.
