Something similar has been occurring with my Medium articles…Admittedly, I don’t check for curation that often because curation does not have the same meaningful distribution that it used to. My articles are doing almost the exactly the same in terms of views/reads/internal reads %, so my guess is it’s part of an A/B test evaluating the feature notation. Prior to August, like many others here, I seemed to strangely being auto curated on 90%+ of my articles, but this had little impact on article performance.