I disagree. A 182cm (6ft) person can weigh only 82 kg (~180lbs) to be not considered overweight. If you have any amount of muscle and don't have a single digit bodyfat percentage, you're easily overweight according to BMI. And it only gets worse when you get taller. Similarly, short people can be seriously fat and not be considered overweight according to BMI. The exponent in the BMI formula should be larger than 2.

I don't know. My husband is 6'4" and is scrawny build but with a bit of extra around the middle and he can feels 'overweight' right around the same time he hits the BMI level. I'm 4'11" and definitely feel the that BMI is correct with how my body feels.

BMI is const x weight/height^2. Volume is proportional to height^3, unless you believe tall people grow in only 2 directions. So BMI ~ height. That's why most of the NBA is obese.

Do humans grow in 3 dimensions? Research suggests the power should be slightly less than 2. e.g. http://www.ncbi.nlm.nih.gov/pubmed/15761809

Some groups of athletes list overweight because muscle is very dense - BMI doesn't distinguish between muscle and fat.

However, this does not mean that BMI is useless even here - overweight athletes often suffer similar medical conditions as overweight people generally.

Everyone mentions athletes whenever BMI is mentioned.

Most people are not athletes. Most people do very little exercise. When using BMI for an individual you ask them if they do any exercise, but for populations it's fine.

I mention the NBA due to height, not athleticism. BMI is an overestimate for tall people and an underestimate for short people, due to the scaling law I described.

Most people are not NBA tall, so luckily it's no such a problem for BMI over mass populations.