We have investigated the effects of Al2O3 facet coatings on the reliability of GaAs‐ (Al,Ga)As double‐heterostructure stripe‐geometry (12 μ) lasers operated cw in a 70 °C dry‐nitrogen ambient with power outputs equal to 5±2 mW/facet. No improvement was seen in the median lifetime of coated devices (some of which also had their sidewalls etched to remove saw‐cut damaged material) in either a nominally short‐lived class (≲100 h, 70 °C) or a long‐lived class (≳1000 h, 70 °C). Corroborating observations of the luminescence originating in the active layer and viewed through a partially contacted n side showed that commonly seen defects within the stripe, thought to be among the causes of laser degradation, were not eliminated by the presence of facet coatings. These defects are (i) a structureless ’’darkening’’ which, when it occurs, always does so at one or both mirrors, and (ii) the previously identified dark‐line defect (DLD) which often, but not always, occurs at a mirror. The DLD is associated more with the short‐lived class and the ’’darkening’’ is associated with the long‐lived class. While neither the suspected origins of laser degradation, particularly those associated with the mirrors (DLD and ’’darkening’’), nor the median lifetimes were altered by the presence of facet coatings, the so‐called facet erosion (which by all appearances is a cosmetic effect in our testing regime) was found to be considerably retarded. This confirmed other accumulated evidence pointing to the lack of any cause‐and‐effect relationship between facet erosion and lifetime. Thus, our devices do not appear to become inoperable at 70 °C because of gradual mirror damage or because of saw- ‐cutting damage to the sidewalls. In addition to suppressing mirror erosion, the facet coatings were also found to present a substantial barrier to degradation‐promoting copper in‐diffusion so that with our reliability results we were able to conclude that uncoated devices do not become inoperative because of the transport of copper to the facets during aging, the gold‐plated copper bonding stud notwithstanding. Modest protection at best was provided by the mirror coatings for devices passively stored in either liquid water at 22 °C or an 85 °C 85% RH environment. Furthermore, not only were the coatings ineffective in preventing the occurrence of either the short‐term or long‐term degradation mechanisms, they were equally ineffective in preventing either of the two known temporally saturable modes of degradation (the previously identified falloff in laser power occurring within the first hundred hours of laser operation and the storage‐aging mechanism). Coating removal experiments did indicate that the temporally saturable modes were reversible, strongly suggesting that they may be two different manifestations of the same surface‐related mechanism. The claims of other workers that facet erosion produces long‐term degradation of lasers in a dry‐nitrogen ambient and that such degradation can be eliminated by the use of facet coatings are not supported by our data.