A new study from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) demonstrates that large language models (LLMs) can be used to automate the planning and execution of complex, multi-step robotics tasks. The system, named ‘Code as Policies,’ allows robots to interpret high-level instructions written in natural language and generate corresponding executable code for physical actions. This approach bypasses the need for traditional, painstakingly hand-coded robotic programs for each specific task. The research shows that LLMs, when provided with a library of simple software primitives, can chain these functions together to perform intricate operations like stacking blocks in a specific pattern or sorting items by color. The method represents a significant step toward more flexible and adaptable robotic systems that can understand and act on general commands. Read the full article at https://news.mit.edu/2023/robots-language-models-code-policies-1004.