Figure’s humanoid robot bests Tesla’s Optimus in latest AI feat

    207

    14th March 2024 – (San Francisco) Figure, an innovative startup in partnership with OpenAI, has escalated the competition in the development of humanoid robots, presenting a new video that showcases capabilities surpassing those of Tesla’s Optimus. Released on Wednesday, the video features Figure 01, the company’s latest iteration of robotic technology, engaging in a seamless, natural interaction with a human.

    The footage, now available on YouTube, captures a strikingly lifelike conversation between a person and Figure 01 (pronounced “Figure One”). During the interaction, the robot identifies objects in its vicinity with remarkable accuracy, stating, “I see a red apple on a plate in the centre of the table, a drying rack with cups and a plate, and you standing nearby with your hand on the table,” all in a tone indistinguishable from a human’s.

    The human participant then requests an edible item, prompting Figure 01 to adeptly hand over the apple. The demonstration continues as the robot flawlessly responds to subsequent requests to pick up trash introduced into the environment, all while explaining its actions with a brief lag in processing time.

    The robot’s final performance is a testament to its precision and dexterity, as it assesses the arrangement of dishes on the table and, upon confirmation, places them accurately in the drying rack. Figure 01’s actions, including the nuanced flipping of a cup and careful positioning of a plate, underscore the advanced stage of development achieved by the startup.

    The video serves as a compelling illustration of Figure’s progress in the realm of artificial intelligence and robotics, inviting viewers to witness the full extent of the robot’s abilities.

    This latest showcase of technological prowess by Figure emerges amidst a backdrop of scepticism regarding Tesla’s Optimus. Despite Elon Musk’s initial unveil in 2021, which featured a human in a robot costume, Tesla’s subsequent demonstrations have been subject to scrutiny. Critiques centre on the authenticity of the robot’s autonomous capabilities, particularly after a close analysis of a video where Optimus folds laundry suggests teleoperation, a method of remote manipulation, may have been at play.

    Contrastingly, Brett Adcock, co-founder of Figure, asserts that their robot’s performance is devoid of such aids. “The video is showing end-to-end neural networks. There is no teleop,” Adcock stated, emphasizing the genuine nature of Figure 01’s interactions and movements.

    The claims of genuine autonomy are bolstered by the credentials of Figure’s team, which includes experts from renowned entities like Boston Dynamics and DeepMind. Corey Lynch, a key player in Figure’s AI development, expressed his pride on social media, highlighting the rapid advancements in the field that have led to the current state of humanoid-robot interaction.