"It's nearly irrelevant for anyone who isn't a couch potato, and even for them, too."
I don't know about that, pretty much every study I read shows increased risk of metabolic syndrome and other negative health consequences as BMI category increases from normal, to overweight, to the various obese categories. BMI for most people is a good measure of "fatness" and that is where the risk probably lies. Fitness and fatness appear to be largely independent risk factors, so it is not accurate to say it is only good for couch potatoes. A fat marathoner is still going to carry the risk of the fatness even if they are not getting the risk of being sedentary.
"I'm ~5'7, 145lbs race weight and 8% bodyfat. That puts me at 22.7 on the BMI scale...I must be fat, right?"
No, that puts you almost dead in the middle of the normal range (20-25), >25 is overweight, >30 obese.
The one group that BMI probably isn't very good for predicting risk is folks who are abnormally muscled for their height, which are mostly young men. So I wouldn't sweat it if I were in that category.