Covariant, a robotics startup, has unveiled its latest innovation, RFM-1, an 8 billion parameter Robotics Foundation Model designed to endow robots with human-like reasoning abilities. Inspired by the capabilities of ChatGPT, RFM-1 aims to revolutionize the way robots interact with the physical world. Leveraging a vast dataset accumulated over four years from Covariant's robots performing picking and placing tasks in warehouses, this model processes images, videos, sensor readings, and more, to predict actions for robots. Covariant's approach represents a significant shift towards creating more adaptable and efficient robots capable of learning and reasoning in a manner similar to humans. The company's efforts have been recognized and covered by major publications such as TechCrunch, WIRED, and The New York Times, highlighting RFM-1's potential impact on the future of robotics. Brian Heater and Lerrel Pinto have notably contributed to the discourse, with Peter Chen, CEO of Covariant, demonstrating the technology's capabilities.