A new study from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) demonstrates that large language models (LLMs) can generate functional, executable computer programs from natural language descriptions. The research team developed a framework called “Program Synthesis with Natural Language” (PSNL), which translates user requests into code by breaking down the problem and verifying the output. In tests, the system successfully created programs for tasks like data analysis and simple game logic directly from conversational prompts. The work highlights progress in bridging human intent and machine execution, though challenges remain in handling complex, multi-step software engineering projects. Read the full article at https://technologyreview.com/2024/05/15/llm-code-generation-study.