From Dr. Robert Lustig in his book, Metabolical (highly recommended):
"Michael Pollan (full disclosure, he’s a friend), in his now-famous New York Times Magazine article, espoused seven simple words: Eat food, not too much, mostly plants. Three separate clauses, but I think that each clause is misleading. Eat food doesn’t take into account that some people may do better on a low-fat diet, while others may do better on a high-fat diet. Not too much doesn’t say how you are supposed to moderate that, as it doesn’t take into account food addiction or what generates satiety. And mostly plants doesn’t take into account that Coke, French fries, and Doritos are all plant-based. If you buy your organic, all-natural, GMO-free tortilla chips at Whole Foods, you’re still stuffing your liver and starving your gut—you’re just paying more for the privilege."
> doesn’t take into account that Coke, French fries, and Doritos are all plant-based
By the very definition Pollan gives, none of those are food, they are lab concoctions that have been invented in the last few generations. Like the advice given about McDonalds by all the doctors and nutritionists in Super Size Me, a healthy person should never eat them.
When Pollan says "eat food", he means "real" things that exist naturally. [Note 1]
He also introduces the "grandma test" which is to say you shouldn't eat anything your grandma wouldn't recognize as food. A glass of black fizzy liquid sugar? no. Drink water.
[Note 1] I often think about this like the periodic table. Many of the elements that have been on the table forever (low atomic numbers) can be found just lying around on the ground (Gold, Copper, etc.). They are "real".
Almost all the ones added in the last few decades (nihonium, moscovium, tennessine) (high atomic numbers) never exist naturally and must be concocted in a lab under extremely specific circumstances. Often they only exist for a fraction of a second. They are "fake".
Can you find an Apple, spinach, meat or fish out in nature? Yes - that's "real" food.
Can you find Doritos or coke out in nature? No - that's not food.
This feels like an overly complex response to a statement that's intentionally simple though.
The whole point of Pollan's statement is to have simple guidelines that ignore optimizations in favor of directionally good flexibility, which makes me feel like Lustig either misses Pollan's point, or just believes in a completely different philosophy.
"Michael Pollan (full disclosure, he’s a friend), in his now-famous New York Times Magazine article, espoused seven simple words: Eat food, not too much, mostly plants. Three separate clauses, but I think that each clause is misleading. Eat food doesn’t take into account that some people may do better on a low-fat diet, while others may do better on a high-fat diet. Not too much doesn’t say how you are supposed to moderate that, as it doesn’t take into account food addiction or what generates satiety. And mostly plants doesn’t take into account that Coke, French fries, and Doritos are all plant-based. If you buy your organic, all-natural, GMO-free tortilla chips at Whole Foods, you’re still stuffing your liver and starving your gut—you’re just paying more for the privilege."