I don’t think that is the point. The argument is usually that we cannot predict what will be “high impact” 20 years from now but the current system works well enough that it is a net benefit despite a lot of research not being directly applicable in the end.