Myeloid cells are critically involved in the pathophysiology of cancers. In the tumor microenvironment (TME), they comprise tumor-associated macrophages (TAMs), neutrophils (TANs), dendritic cells, and myeloid-derived suppressor cells, which are further subdivided into a monocytic subset and a granulocytic subset. Some of these myeloid cells, in particular TAMs and TANs, are divided into type 1 or type 2 cells, according to the paradigm of T helper type 1 or type 2 cells. Type 1-activated cells are generally characterized as cells that aid tumor rejection, while all other myeloid cells are shown to favor tumor progression. Moreover, these cells are often at the basis of resistance to various therapies. Much research has been devoted to study the biology of myeloid cells. This endeavor has proven to be challenging, as the markers used to categorize myeloid cells in the TME are not restricted to particular subsets. Also from a functional and metabolic point of view, myeloid cells share many features. Finally, myeloid cells are endowed with a certain level of plasticity, which further complicates studying them outside their environment. In this article, we challenge the exclusive use of cell markers to unambiguously identify myeloid cell subsets in the TME. We further propose to divide myeloid cells into myeloid regulatory or stimulatory cells according to their pro- or antitumor function, because we contend that for therapeutic purposes it is not targeting the cell subsets but rather targeting their protumor traits; hence, myeloid regulatory cells will push antitumor immunotherapy to the next level.