This usually has little to do with the field of nutrition or the studies themselves. It has to do with how the media reports the findings of the studies, usually without properly understanding them and without putting them into context. Everything has to be revolutionary, has to prove everything else wrong, even though it's extremely difficult to perform a valid scientific study that proves anything wrong.

My wife has a long-time friend who became a nutritionist, and she'll tell anyone that will listen that most people have a pretty good idea of how to eat right if they don't over-think it, but it becomes very difficult to put it into practice, because everything we do to make food convenient tends to also make it bad for us. Another issue is that medical doctors don't take nutritionists seriously (despite medicine also being a constantly-shifting landscape), and rarely pay attention to diet for anything other than controlling your weight, blood sugar, and cholesterol. This is why we end up with things like "Irritable Bowel Syndrome", which, for most people is likely an undiagnosed food intolerance (and given the most common ingredients in processed foods, it's probably soy or lactose).

How do you know the studies say those things if you don't read the studies?

Is nutrition bunk, or is reporting on nutrition just very bad?

There's a bunch of things that make nutrition science weird:

1) everyone feels free to talk about it as if they know what they're talking about. You see this on HN. You see this on any forum. But you also have frauds like "Dr" Gillian McKieth (once popular TV / book nutritionist) who put out vast quantities of bullshit information.

2) huge multinational companies engage in well-funded dis-information campaigns to discredit scientists. They do this to avoid increased prices through taxation.

3) science reporting is almost entirely terrible. Those studies don't usually make these grandiose claims. Idiot news paper reporters make the grand claims.

4) researchers can't run real experiments. Most of the science comes from self-reported uncontrolled case studies. Of course it's not going to be as good as a meta-analysis of randomised controlled trials.