A field guide to the 'world model' buzzword explosion

What it does This repository is a curated bibliography that stitches together prior awesome-lists for autonomous driving and robotics into a broader index of world-model research. It sorts papers, surveys, and blog posts by domain—embodied AI, game simulation, science, NLP, and autonomous driving—tagging each entry with quick-link badges for arXiv, code, and project websites. The maintainers explicitly aim to bridge siloed communities and track a field where the term “world model” is, by their own admission, applied with “varying definitions across different domains.”

The interesting bit Instead of narrowing the definition, the repo leans into the ambiguity, treating “world model” as an umbrella large enough to cover Ha & Schmidhuber’s 2018 generative environments, LeCun’s autonomous-intelligence roadmap, and modern video-generation surveys alike. That inclusiveness is either a public service or a quiet surrender to buzzword inflation, depending on your cynicism level.

Key highlights

• Cross-domain coverage spanning embodied AI, autonomous driving, game simulation, NLP, science, and theory.
• Every entry carries quick-access badges linking to arXiv papers, project websites, and code repositories.
• Builds directly on two prior specialized lists (robotics and driving) and extends them into a wider map.
• Released under CC0, making the bibliography effectively public-domain fodder.
• The project claims it reached 1,000 GitHub stars within 30 days of launch.

Caveats

• The list is largely aggregation: entries are titles and links with minimal critical commentary or synthesis.
• Some “News” timestamps use future dates (e.g., 2025-11), suggesting the changelog may be aspirational or copy-pasted.
• Starred [⭐️] entries hint at editorial picks, but the selection criteria are never explained.

Verdict Worth bookmarking if you are a researcher or practitioner drowning in world-model papers and need a quick, domain-sorted index. Skip it if you want deep synthesis, implementation details, or a rigorous gatekeeping of what actually counts as a world model.