Studies over the past decade have shown that the initial mass function (IMF) is the same for massive stars born in the OB associations of the LMC and SMC as in the associations of the Milky Way: the slope of the IMF is essentially Salpeter (Γ ~ −1.3), despite the factor of 4 difference in metallicity between these systems, and despite a factor of several hundred in stellar density between the sparsest and richest OB associations. However, there does appear to be a number of massive stars that are born in relative isolation, and the IMF of this mixed-age, field population is quite different than that of OB associations, with Γ ~ −4 in all three galaxies. The distribution of stars in the HR diagram is in excellent agreement with the Geneva group's evolutionary models for stars with masses with no “main-sequence widening problem” left to be solved. The massive stars born in clusters are formed quite coevally (Δτ < 1–2Myr), which allows us to use the “turn-off masses” to determine what mass objects become Wolf-Rayet stars of various types, and new results are briefly described. For the LMC, WNEs come from a wide range of masses, WCs come only from the highest mass stars, and Ofpe/WN9 “slash” stars come from lower mass OBs. Recent work on the R136a cluster (described in Hunter's review talk) suggest that there is no such thing as an upper mass cutoff to the IMF, at least not one that has been found observationally: for the youngest clusters (2 Myr and younger), the mass of the highest mass star present is simply dependent upon how populous the cluster is; i.e., the IMF is truncated by statistics, not physics.