as far as breeding something more noticeable is probably possible, the problem is when you find that trait and breed it into the variety, do you bring down the productivity of the variety. Is a 2% reduction in productivity of the chicken (over the course of its life) pay for the 3-4 seconds it takes for a chicken sexer to make a decision? Even if the sexer makes $60k/year and it takes them 4 seconds to make a decision, it only costs 3 cents to make a decision. There are a whole bunch of other considerations (life of avg layer, input costs per egg, etc). I don't know the answer, but is all that worth the risk that the productivity takes a 5% or 10% hit instead of 2%.
In cattle, this is a lot easier, because bull semen can be diluted quite heavily and still produce viable offspring.