According to a new Working Paper by the Institute for Fiscal Studies, titled " Technology, skills, and performance: the case of robots in surgery " (freely available here ) robots have the potential to reduce variation in patient outcomes that would arise from heterogeneity in surgeons’ skills. Interestingly, it also points out an important inefficiency of our innovation systems: lower skilled surgeons may face a higher cost of using the robot as they are usually operating in less "ready" settings, and thus, despite them/their patients having the highest gains, they are less likely to use the robot on any given patient. These findings, that we would love to see replicated in other ecosystems (do we have connections that would consider working on this together?) seem to highlight the existence of an organisational gap between the necessity of testing new technologies (R&D) in high performing settings to better control confounders, and the need to transfer those n