Covariant Unveils RFM-1: Language Model for Industrial Robots

Covariant, a startup specializing in AI-powered warehouse automation systems, has unveiled its latest innovation – the RFM-1 language model. This model allows industrial robots to ask for instructions when they encounter tasks they can’t complete on their own, revolutionizing the way robots interact with their environment.

The RFM-1 model, with its 8 billion parameters, is trained on real-world data captured by warehouse robots in production environments. This sets it apart from other AI models that are trained on lab footage, making it a more valuable source of knowledge for the robots it powers.

One of the key features of RFM-1 is its ability to process natural language instructions, enabling workers to teach robots new tasks using plain English prompts. This eliminates the need for writing custom code, saving time and effort in programming warehouse automation systems.

Moreover, RFM-1 can generate short videos depicting robots performing tasks, allowing them to plan their actions and make decisions in real-time. This capability enhances the efficiency and flexibility of warehouse automation systems, making them more adaptable to changing tasks and environments.

Covariant plans to roll out RFM-1 to customers’ warehouse robots in the coming months, with a focus on creating more advanced AI models for a broader range of robot configuration tasks. By increasing the amount of training data collected, the company aims to accelerate the development of cutting-edge automation solutions for the industry.

With the support of investors like Index Ventures and a strong commitment to innovation, Covariant is poised to lead the way in transforming the future of warehouse automation. Stay tuned for more updates on this exciting technology and its impact on the robotics industry.