The GPT-enhanced humanoid speaks and thinks

The Figure has unveiled the initial outcome of its partnership with OpenAI, showcasing advancements in humanoid robot capabilities. The Figure 01 bot engages in live conversations in a recently published video.

The progress made in the development of the Figure is exceptional. After assembling a team of experts from Boston Dynamics, Tesla, Google DeepMind, and Archer Aviation, Entrepreneur Brett Adcock came out of stealth mode only last year. Together, their goal is to create the world's first commercially viable general-purpose humanoid robot.

Figure 01 was already operational in October and carried out simple autonomous tasks. With its fully developed watch-and-learn capabilities, the robot was prepared to join the BMW workforce in mid-January.

Last month, Figure had an exciting opportunity to showcase its work; the company announced the successful completion of its Series B funding round and a promising collaboration agreement with OpenAI. Together, they will pioneer the development of next-generation AI models for humanoid robots.

According to Adcock's post on X, it has been confirmed that Figure 01's integrated cameras transmit data to a sophisticated vision-language model developed by OpenAI. The Figure's neural networks can also capture images at 10 Hz using the robot's built-in cameras. OpenAI is also behind the development of speech recognition technology, which enables the understanding of spoken words. The Figure's neural network then processes the information gathered to generate quick and precise robot movements.

Adcock confirmed that the robot was not controlled remotely during the demonstration, and the video accurately captured the real-time speed. This partnership has already achieved something remarkable in just fewer than two weeks. Adcock expressed their ambitious goal of training a world model to operate humanoid robots on a massive scale, reaching billions of units. It is an impressive feat for such a young collaboration. We will not have to wait much longer based on the current pace.