Louis Abraham grouped GitHub PRs since 2025 into 10 clusters of vocabulary using KL-divergence k-means and made a cool dashboard to explore the data. One of the clusters appeared in 2026 and represented 40% of all human-attributed pull requests last month. Its most representative words should look familiar to anyone who uses coding agents.
So: what if one sets up a skill that forbids the top words of this cluster? Is the Claudism solved then?
— Felix