Researchers have introduced a novel framework called GaP, a graph-as-policy multi-agent self-learning harness, to tackle variational automation tasks. GaP aims to combine the interpretability of robot programming with the adaptability of model-free policies, enabling robots to operate reliably in commercial and industrial settings. This approach focuses on variational automation, a class of tasks characterized by significant variations in object geometry and pose. By leveraging GaP, robots can learn to adapt to new situations without requiring explicit programming, thereby enhancing their autonomy and flexibility. The development of GaP has significant implications for the field of robotics, as it can potentially improve the efficiency and effectiveness of automated systems in dynamic environments. So what matters to practitioners is that GaP's ability to learn and adapt in complex scenarios can raise the bar for autonomous systems in industrial applications1.