[sseagull]

As a society, we want to have a "portfolio" of research. We want short-term research with tangible results, moderate-term research that builds on the current state of the art and advances it, and long-term research that might pay off decades from now. You also want to mix in "moonshot" research that likely won't pay off, but if it did would be revolutionary. The proportion you invest in each is up for debate.

Generally, as the time horizon increases, risk increases, and it becomes less and less likely any particular research direction will pay off, but you still need to do it or you won't have anything after all the current directions have dried up, matured, or otherwise run their course.

I think the error people make is to assume academia is supposed to cover all three. Largely, academia covers long-term and maybe some moderate-term and moonshot. Short term is better handled by industry where there is a more direct measure of applicability ($$). But business is generally not interested in the longer term stuff because it is too risky.

So academia as a whole is about forming the bricks that industry uses to build useful things. While any particular paper is not going to be useful to society anytime soon (if at all), the aggregate knowledge will likely be useful at some point in the future.