Awesome repositories are no longer awesome

GitHub’s “awesome” repositories collect curated links and resources for particular fields, but their usefulness is declining for two AI-related reasons. First, lists that are not regularly maintained grow stale: the author points to awesome-R, last updated more than a year earlier, whose resources include a ranking of the most-downloaded R packages from 2016. For someone seeking current guidance, asking a capable AI model may now be more useful than relying on an old directory. These lists may retain value as indexes that automated agents can consult and verify, even if they are less helpful to people.

Second, AI-generated projects and outreach are overwhelming some lists with low-quality submissions. The author describes a proposed addition to a statistics resource collection: StatOrigin, an “open database” of industry statistics. Checking the site and the contributor’s activity revealed at least six vibe-coded websites and more than 100 pull requests promoting them to “awesome” lists. Some repositories accepted similar submissions, while another rejected one after finding inadequate evidence of expert review, independent validation, or learning outcomes. The argument is not that every curated list is worthless, but that they are increasingly unreliable as a general way to discover good resources. As AI makes it easier both to generate tools and to submit them at scale, maintainers face more noise, and the lists’ value depends increasingly on careful human vetting.