We know a lot about the basic physics of energy conservation, chemistry and biochemistry to construct a framework of metabolism and how people use energy and create work, but the specifics that describe exactly what your diet should be if you want to lose weight and keep it off are not known. Likewise we know Newtonian physics well, but kinesiology is not well enough known to describe exactly how walking or running works.
People have an expectation that science, which often knows the fundamental bits of science in some areas, is able to explain all of the derivative bits associated with it - but the world doesn't work that way. Often unraveling the bits and pieces takes significant work and often little progress has been made.
This piece on running hints at that - people are looking for "the best way" even though the basics aren't known. As a result they seem to latch onto anything. Some of the advice is good and some isn't. The same is true for dieting. The basic principals are known, but getting it right at the individual level can be extremely difficult (for example: even in calorie counting the approximations used can often be off by 5% and the regulation of the efficiency of the absorption of food is not extremely well-understood). People are looking for easy explanations and results and, as a result, will try tactics that may have undesirable side effects in order to achieve the core goal.
There are many areas where the knowledge is imperfect, but good enough. Driving the average car at 60 mph may or may not deliver better fuel mileage than at 55 mph - but driving at 70 mph always is worse than 55 mph. Our current rate of adding carbon to the atmosphere will certainly increase global warming although the exact impact on the weather next year is unknown. It is important to understand confidence levels associated with the science and then good enough advice can be given. That message seems to be lost.
Science, particularly fundamental science, tends to built on itself and established results rather than throwing out the old. Relativity and Quantum Mechanics both extend Newtonian physics into new areas - the very fast and the very small - but for all practical purposes Newtonian physics is just dandy for most of the world at our human scale. Moving to complexity gets a bit more muddled as the understanding may not be as deep and calculation may be extremely difficult, but that is what understanding confidence levels is all about.