Researchers at the University of Pennsylvania have developed VLMgineer, an AI-powered framework that allows robots to design, test, and improve tools for completing physical tasks. Instead of directly controlling a robot, the system analyzes a problem, proposes tool designs, simulates their effectiveness, and can even create physical versions through technologies like 3D printing. The AI continuously learns from real-world results, mimicking how humans solve problems through observation and experimentation. Researchers believe this technology could eventually help robots adapt to a wide range of applications in manufacturing, laboratories, agriculture, and everyday life.
