Hugging Face’s acquisition of Pollen Robotics aims to democratize robotics by making AI robot technology more accessible to developers and researchers.
Hugging Face, a company that hosts open source artificial intelligence models and software, has announced the acquisition of Pollen Robotics, the French startup behind the bug-eyed, two-armed, humanoid robot called Reachy 2. This move aims to ‘democratize’ robotics by making its technology more accessible to developers and researchers.
Open source robotics refers to the practice of releasing designs, component details, and 3D models that allow pieces to be manufactured more easily. ‘Hopefully open source can unleash a wide and diverse range of [new robot] capabilities,’ says Clément Delangue, chief executive of Hugging Face. The company plans to sell the Reachy 2 robot and allow developers to download, modify, and suggest improvements to its code.
Open source robotics refers to the development and distribution of robotic systems, hardware, and software under open-source licenses.
This approach allows developers to access, modify, and share robotic technologies freely, fostering collaboration and innovation.
Open source robotics enables researchers, hobbyists, and industries to create customized robots for various applications, from industrial automation to healthcare.
The benefits include reduced costs, accelerated development, and improved adaptability.
Open sourcing hardware has a similar effect as open sourcing software. Robot developers ‘can [3D] print a part if something is broken,’ says Matthieu Lapeyre, cofounder and CEO of Pollen Robotics. This approach also allows for greater transparency and accountability, making it harder to hide flaws or manipulate results.
The hype surrounding humanoid robots has led to some dubious claims. Some companies have posted demo videos on social media that seem to promise incredible abilities, but these claims can be misleading. A system that seems extraordinary online could in fact be teleoperated by a person off camera; it could also fail if conditions change even slightly; or be unable to complete a task reliably.
Humanoid robots are machines designed to resemble humans in appearance and movement.
They typically have a human-like body structure, including a head, torso, arms, and legs.
The first humanoid robot, Wabot-1, was developed in 1974 by the Japanese robotics company 'Waseda University'.
Since then, humanoid robots have evolved to perform various tasks, such as assistance, entertainment, and research.
They are equipped with advanced sensors, artificial intelligence, and robotic limbs, enabling them to interact with humans in a more natural way.
Some robotics researchers, especially those in academia, favor the open approach. ‘Making robotics more accessible increases the velocity with which technology advances,’ says Sergey Levine, an assistant professor at the University of California, Berkeley, and cofounder of Physical Intelligence, a startup working to develop vastly more capable and general purpose robotic models.
Physical Intelligence made the first of its robot foundation models, Pi0, available on Hugging Face in February. The model allows a range of different robots to learn to do a variety of physical tasks. Levine says that researchers in academia and industry have already contributed valuable ideas and tweaks to his product.
Robot foundation models are pre-trained artificial intelligence (AI) models that serve as the building blocks for various robotic applications.
These models learn general-purpose representations of data, enabling robots to understand and interact with their environment more effectively.
They can be fine-tuned for specific tasks, such as object recognition or navigation, without requiring extensive retraining from scratch.
Robot foundation models have been shown to improve the efficiency and accuracy of robot learning, paving the way for more advanced robotic systems in industries like manufacturing and healthcare.
The open approach appears to be gaining momentum across the AI industry. Meta was the first major AI company to offer a cutting edge open weight model when it released Llama in 2023. Several other cutting edge open weight models have followed, including DeepSeek and an upcoming open weight model from OpenAI.
This trend suggests that open source robotics is on the horizon, with companies like Hugging Face leading the charge. As the technology advances, we can expect to see more accessible and affordable robots that are capable of performing a wide range of tasks.
