The progenitors of SN 2008S and the 2008 luminous transient in NGC 300 were deeply dust-enshrouded massive stars, with extremely red mid-infrared (MIR) colors and relatively low bolometric luminosities (≈5 × 10 4 L ). The transients were optically faint compared to normal core-collapse supernovae (ccSNe), with peak absolute visual magnitudes of −13 M V −15, and their spectra exhibit narrow Balmer and [Ca ii] emission lines. These events are unique among transient-progenitor pairs and hence constitute a new class. Additional members of this class may include the M85 transient, SN 1999bw, 2002bu, and others. Whether they are true supernovae or bright massive-star eruptions, we argue that their rate is of order ∼20% of the ccSN rate in star-forming galaxies. This fact is remarkable in light of the observation that a very small fraction of all massive stars in any one galaxy, at any moment, have the infrared colors of the progenitors of SN 2008S and the NGC 300 transient. We show this by extracting MIR and optical luminosity, color, and variability properties of massive stars in M33 using archival imaging. We find that the fraction of massive stars with colors consistent with the progenitors of SN 2008S and the NGC 300 transient is 10 −4 . In fact, only 10 similar objects exist in M33 (and perhaps 1)-all of which lie at the luminous red extremum of the asymptotic giant branch sequence. That these transients are simultaneously relatively common with respect to supernovae, while their progenitors are remarkably rare compared to the massive star population, implies that the dust-enshrouded phase is a short-lived phenomenon in the lives of many massive stars. This shrouded epoch can occur only in the last 10 4 yr before explosion, be it death or merely eruption. We discuss the implications of this finding for the evolution and census of "low-mass" massive stars (i.e., ∼8-12 M ), and we connect it with theoretical discussions of electron-capture supernovae (ecSNe) near this mass range. Other potential mechanisms, including the explosive birth of massive white dwarfs and massive star outbursts, are also discussed. A systematic census with (warm) Spitzer of galaxies in the local universe (D 10 Mpc) for analogous progenitors would significantly improve our knowledge of this channel to massive stellar explosions, and potentially to others with obscured progenitors.