The vast majority of YouTube videos never become popular, languishing in obscurity with few views, no likes, and no comments. We use information theoretical measures based on entropy to examine how time series distributions of common measures of popularity in videos from YouTube's "Trending videos" and "Most recent" video feeds relate to the theoretical concept of attention. While most of the videos in the "Most recent" feed are never popular, some 20% of them have distributions of attention metrics and measures of entropy that are similar to distributions for "Trending videos". We analyze how the 20% of "Most recent" videos that become somewhat popular differ from the 80% that do not, then compare these popular "Most recent" videos to different subsets of "Trending videos" to try to characterize and compare the attention each receives.