Maybe he's got a finer point, but it actually just looks like an argument against categorization. It's like saying I don't trust math about "triangles" or "scalene triangles" or "rhombuses" when you find out about the special properties of the equilateral triangle or the square.

The fact that there are differences between elements of a category does not eliminate the utility of the commonalities shared by elements of a category. It does limit that utility, yes.

For example, just because you are getting plenty of protein, if you somehow completely avoid one of the amino acids that the human body uses but can't synthesize, then eventually you will have some fairly specific health problems. That's not strong evidence that it's worth micromanaging your macronutrients by tracking your intake of all amino acids individually. (It might be; I haven't seen it studies either way.)

Maybe I'm missing some context, but I also get the "anti-science" vibe from the image.