A generalizable robot is a robot that can carry out a wide range of tasks in different environments, as opposed to a machine built and programmed for one job. A conventional industrial robot repeats a fixed motion in a controlled cell. A generalizable robot is meant to handle unfamiliar objects, follow instructions given in ordinary language and learn new tasks without being re-engineered. The idea is closely tied to advances in artificial intelligence, in particular models trained on large sets of images, text and recorded movements.
The concept is a leading theme in robotics investing, and the humanoid robot is its most visible form. The commercial argument is that one machine sold in large numbers for many uses could be cheaper than many specialised ones, extending automation into warehouses, homes and services. The obstacles are reliability, since error tolerance is low in real settings, the scarcity of training data and cost. Most systems remain at pilot stage.