You are characterizing my argument unfairly. I have always argued that directed selection is prudent and profitable. Directed selection requires quantification of traits, saturation and isolation. Directed selection is a long-term, large-scale and professional endeavor.
What I have argued against is the almost magical belief that starting a microscopic apiary in an unbounded, out-crossing population and simply adopting a live-and-let die management style will result a new lineage of perfected bees. Modern breeding doesn't work effectively in that milieu, especially for bees that have evolved multiple behaviors to ensure genotype conservation across millions of years.
To respond to your specific complaint:
Eggs that are homozygous at the "csd" gene develop into diploid drones, these are detected and killed by nurse bees (in most cases). (That's blind science-speak -- eggs that have identical "sex-determiner" alleles on the egg-gamete and the sperm will when fertilized express lethally). Sex determination in bees has nothing to do with the mammalian XY chromosome system.
This means a high proportion of brood developing from a queen that has mated with a nest-mate drone will be lethal (and is the source of the observed condition called "shot-brood". Bee colonies cannot be inbred -- these will die out. Inbreeding includes daughter colonies (which will converge to lower variance) as well as the foundress colony. The csd gene has 16-20 polymorphisms that are cross-compatible. The queen lineage contributes two possible alleles of these 20, and the drone mated with the queen contributes one. If the drone csd allele exactly matches one of the two queen alleles -- 50% of the produced eggs will die. In the virgin queen mates with her brother scenario --- alleles will match (in some proportion) and eggs will die. Locally in-bred populations are at risk for a toxic spiral of less-and-less fecund matings.
What this means, of course, is that out-crossed colonies have much higher fitness. Bees, in nature, will constantly reconverge on the normative genotype. This is why they form a single interbreeding species, and not a thousand local, incompatible taxa. I've explained before the plant-insect co-evolutionary logic for this conservatism. Its a bit of a trap for the bees, but once engaged difficult to shift the overarching direction.
The US Russian Queen breeder program was designed with bee genetics in mind. Colonies were assigned to one of three groups. The groups were physically rotated. A queen grafting source (measured for improved expression) was physically moved to a deliberately isolated out-yard surrounded by drone hives moved from a distant location. Remember my preconditions: Quantification, Saturation and Isolation.
The Russian queen program appears to have fallen apart as the scientists withdrew funding/support. I have no idea if commercial interest, personality conflict, or genotype unsuitability was the driving factor in the programs change of character.