This... our understanding of biology is way too primitive to have a meaningful mechanistic understanding of what is healthy, and what is not. Most of the nutrition advice is based on simple observational correlations that are assuming a cause and effect that just isn't there. People with high cholesterol also tend to have more cardiovascular disease, but it turns out eating cholesterol and fat doesn't actually increase risk of cardiovascular disease. People who eat a lot of fish tend to have better health outcomes, but it turns out taking fish oil does not reproduce those outcomes, and so on and so on.

Scientific nutrition is mostly just "scientism" - an irrational overconfidence bordering on a religious faith in unfounded assumptions based on observational studies, without admitting what we don't know.

I think it is reasonable to avoid trying to make decisions about diet based on this stuff, but I think ideas like the paleo diet or evolutionary nutrition make a lot of sense- eat diets similar to those that humans have eaten safely for a long time, as those are what we are likely adapted to. Interestingly though this itself is massively diverse: there are hunter gatherer societies with almost every diet composition imaginable: from artic diets that are high protein and fat but nearly zero carb, to cultures like the kitavans whose diet is very high carb and low protein. Our metabolism is very adaptable and any diet that is mostly fresh nutrient dense foods from plant and/or animal sources is probably about equally healthy.

Ironically, the stress of worrying constantly about if your food is optimally healthy, is probably more harmful to your health than anything typically considered unhealthy.