ICLR 2026 – Institutional Affiliations Dataset and Analysis

A tiny GitHub repository about who is attached to papers submitted to ICLR 2026 — a major artificial intelligence research event — somehow turned into exactly the kind of internet spectacle people can’t resist. On paper, it’s just a dataset and a few scripts for scraping public listings, downloading papers, reading author affiliations, and turning the whole thing into charts. In the comments, though? Absolute blood sport.

The strongest reaction was a familiar one: people arguing over whether counting company and university names is a public service or a weird obsession. One camp treated it like a transparency win, saying the community deserves to know which labs, big tech firms, and elite schools dominate the conversation. The other side rolled its eyes and called it leaderboard brain, warning that affiliation charts can flatten messy reality into a scoreboard for clout-chasing.

Then came the classic academia-vs-industry drama. Some commenters joked that every conference chart eventually becomes “surprise, it’s the same five logos again”, while others defended the work as a useful reality check for anyone wondering where the money and influence are. The humor was predictably sharp: memes about “corporate bingo cards,” jokes that the real machine learning benchmark is now counting badges on author pages, and quips that researchers can’t even submit a paper without someone turning it into a treemap. In short, the repo is simple — but the reaction says a lot about the community’s anxiety over power, prestige, and who gets to shape the future of AI.