Frameworks and reference material for thinking about scientific work itself. The current focus is on a seven-vector taxonomy for classifying AI and physical sciences papers by their dominant contribution type (Method, Theory, Resource, Systematization, Position, Discovery, or Application). This kind of meta-scientific scaffolding is useful when navigating a large literature: knowing what kind of claim a paper is making changes how you read it and what questions to ask.

PaperKey Idea
AI & Physical Sciences Taxonomy: A Seven-Vector FrameworkWorking taxonomy for classifying papers as Method, Theory, Resource, Systematization, Position, Discovery, or Application