The primary issue they are pointing out is that most nutrition research is based on epidemiological studies, not intervention studies. In an epidemiological study, you don't do anything, you observe and record data. Then you look for correlations. However, correlation cannot prove causation. There is a strong correlation between the presence of firemen and fires, but firemen don't cause fires. There are other statistical issues like "P-hacking" and some types of bias as well. The big problem is that nutrition policy, and the dietary guidelines, are based on these weak and faulty studies. Nobody is doing intervention trials, most notably the randomized control trial, which is the gold standard for research. The issue is that nobody wants to fund them because they are so expensive. You know how hard it is for people to stick with a diet. Imagine how hard it is to do diet research! Compliance can be bought but it costs too much. Our whole foundation of nutrition knowledge is built on shaky ground. The best I can say with certainty is to eat real food.