Humanoid robotics startup, Figure AI, which has also received investment from OpenAI, has just unveiled its new project, Figure 02. This robot is the first of its kind to integrate OpenAI’s GPT artificial intelligence model, the same one that makes ChatGPT work.
About Figure AI
Figure AI is not the usual startup. Among its main backers are giants such as Microsoft, Nvidia, and OpenAI. The company focuses exclusively on developing humanoid robots, starting with a simple but fundamental belief: the world was created by humans and for humans, so it is robots that must adapt to the world, not the other way around.
The company is particularly convinced that robots must adapt to the world of work, factories, logistics warehouses, and stores. There are currently 10 million undesirable jobs in the United States, where workers are exposed to a high risk of accidents. In addition, there are 10 million vacancies, including 7 million in the logistics, transportation, and retail sectors, but only 6 million workers potentially available to fill them.
Figure AI’s idea, therefore, is to create robots that can replace humans in all these roles, filling the gap in the labor market.
About Figure 02
The robot can understand and analyze images, including real-time streams of its surroundings, and can also respond and communicate with voice, using natural language. The main innovation of OpenAI’s newest GPT-4o (“omnia”) model is precisely its ability to generate multimodal content, switching from text to images and audio.
The Figure 02 is distinguished from the previous model by a sleeker and more functional design, featuring a matte black finish to replace the chrome and full integration of cables within the limbs. These aesthetic improvements are not just for beauty, but help make the robot more robust and reliable.
At the heart of the humanoid is an enhanced AI computation and inference unit, with three times the capacity of the previous version, which enables Figure 02 to perform complex tasks autonomously. With six cameras and a sophisticated vision and language model, the robot can perceive the environment and interact intelligently. Another distinctive feature of Figure 02 are its hands, which have 16 degrees of freedom and can lift weights of up to 25 kg.