The author's point was that this dynamic does not apply to managerial positions. They are, in their argument, almost entirely composed of "non-heroes" and thei

The author's point was that this dynamic does not apply to managerial positions. They are, in their argument, almost entirely composed of "non-heroes" and their actual work is of speculative value. Hence why they are suggesting that 50% of them can be randomly removed without a corresponding loss of effective productivity. I don't know if that's true or not (though I'd love to see it studied), but I can at least see that logic behind the claim - roughly summarized as: "The ability of a manager to increase the productivity of workers is inversely proportional to the complexity of the work being done."