It is in fact not at all reasonable. They are saying that the BLS stats can't be trusted because they totally misunderstand the survey methodology. That isn't a reason!

I’d counter that if we were doing a good job gathering data that these structural biases could be compensated for with more conservative initial numbers.

At some point a lack of decision to take compensating action becomes faking the numbers.