The declining marginal return of research and development

The decreasing benefits from specialized, derivative work, viewed from the perspective of the overall history of science, are acquired at substantially greater cost. The costs to societies of early support of science tended to be minimal. Generally, as in the ancient Mediterranean or Medieval Europe, it consisted of little more than the support of individual naturalists or mathematicians and their students, or the support of religious specialists who also performed scientific inquiry. Science today, in contrast, is a costly process involving complex institutions, sophisticated technology, and large, interdisciplinary research teams.

This costly science certainly produces astonishing results, but these cannot be claimed to be more valuable than the generalized knowledge of earlier, less expensive science. As impressive, for example, as modern travel technology is, it is hard to argue that it is of greater consequence than the development of the wheel, or of water craft, or of the steam engine. As astounding as it is to put human beings on the moon, this is not of greater import than the principles of geometry or the theory of gravity. However valuable may be genetic engineering, the benefits of this complex process must always be attributed in part to the nearly cost-free work of Gregor Mendel. (...). Exponential growth in the size and costliness of science, in fact, is simply necessary to maintain a constant rate of progress.

Quoted from: "The Collapse of Complex Societies", Joseph A. Tainter, 1988 (Amazon link). Excerpts.