You're technically correct. It's a formula, it's never 'wrong'. The issue is the mapping of the formula to health outcomes and categories and the limitations of it as a tool.
I wouldn’t put “denser” people in the same group as tall people. Having heavy bones is an old excuse for being skinny-fat which is unhealthy
But at a population level (as the article uses it) it’s mostly fine. Unless the sample skews tall or muscular like the Dutch or bodybuilders.
BMI is a screening tool with high sensitivity and low specificity. It is useful for selecting patients for further, more specific tests, such as waist measurement.
But when talking about how the average BMI of all Americans has gone up, it’s mostly fine.