In the 1980s and 1990s, researchers came to understand poor urban neighborhoods as blighted, depopulated areas, based on important ethnographic observations in a handful of cities. This image helped inform influential theories of social isolation and de‐institutionalization. However, few scholars have examined whether those observations were representative of poor neighborhoods nationwide—and whether they are representative today. Based on a descriptive analysis of the largest 100 U.S. metropolitan areas using normalized census tract boundaries, we document an important transformation in the conditions of poor neighborhoods. We find that the depopulation in poor neighborhoods often reported in cities such as Chicago and Baltimore was, in fact, typical across cities in 1990. Today, it is not. Moreover, heterogeneity across cities has increased: The experience of neighborhood poverty is likely to depend more today than in 1990 on the city in question. In fact, the most typically studied cities, such as Chicago, Baltimore, Philadelphia, and Milwaukee, are increasingly atypical in this respect. Addressing today's core questions about neighborhood effects, how and why they matter, requires paying far greater attention to heterogeneity, conducting more ethnographic observation in ostensibly unconventional cities, and addressing the historically extreme conditions in a newly unique subset of cities.