OK, first, that sentence
does not mean "smaller things are harder to click or tap". That sentence would still hold true if smaller things were
easier to click or tap. Besides, in 1954, when Fitts' paper appeared, it was already understood that this was the case.
I'm not being overly pedantic here -- Archimedes' principle is not that "whether an object sinks or floats depends on what it's made of", that had been known since the dawn of time. Archimedes' contribution was to give it a quantitative formulation, which is what we now call Archimedes' principle -- the force is equal to the weight of the volume of liquid that the immersed object displaces. Similarly, Fitts' law is not that larger targets are easier to hit, that has been understood forever; it's quantifying this statement, and the end result is what we now know as Fitts' law.
Second, the implication "make bigger buttons" is not at all valid in every circumstance -- which is why I warmly suggest reading the actual paper to anyone who cites it.
Fitts' work applies to repetitive movements of equal amplitude and speed, not necessarily made with active cognitive involvement, and assumes that targets are known in advance and -- although not explicitly stated, but implicit in the experiment -- easily distinguishable. It will tell you nothing useful about the size of buttons in dialog boxes, for example. (Edit:) this blind interpretation would also suggest, for example, that making the macOS menu thicker would help, but it doesn't; or that simply making buttons thin, but placing them on top of the screen, would also help -- which holds is only partially true for mouse-driven interfaces, and not at all for touch interfaces.