As someone interested in what applied physics can tell us about the capabilities of a more fully developed technology base, I would take issue with Mill here:

We cannot, indeed, foresee to what extent the modes of production may be altered, or the productiveness of labour increased, by future extensions of our knowledge of the laws of nature, suggesting new processes of industry of which we have at present no conception.

This statement falls into a pattern still familiar today -- the confusion of science (which advances by learning new things) with technology (which advances by making new things). One cannot predict what one will learn, simply because to do so would require that it already be known. Predicting (more accurately, inferring) that something can be made, in contrast, is routine in technology development. No new laws of nature are involved in developing a next-generation aircraft, merely the working out of the intricate consequences of intricate systems based on materials and principles that are already known. Well, give or take some fix-up and debugging.

In practice, of course, many of the most interesting advances in technology result when science discovers new materials and phenomena, and many of the most interesting advances in science rely on technology to make new instruments and tools. The conceptual distinction remains important, however, even when the activities are intimately intermixed within a single research group.

Because technology need not rely on new knowledge of nature, applied physics can be used to determine some facts about as-yet-unimplemented revolutionary technologies. These facts typically tell us more about the lower bounds on future achievements because, although careful inference can establish possibilities, it cannot so easily establish impossibilities (unless some natural law would obviously be violated -- perpetual motion, and all that). Describing upper limits on what can be done is naturally more difficult because it would require a proof of the unworkability of an unbounded set of possible designs, and because science may turn up new, exploitable materials and phenomena. Describing lower limits on what can be done is possible because this requires only that particular, high-margin-of-safety designs will work as physics says they will.

This knowledge regarding feasible developments is of first-rank importance to policy development, yet is commonly ignored.

Words and ideas I offer here may be used freely and without attribution.