Research

Research lives or dies by what you can find again. The papers I bookmarked but never read, the perfect quote I remember reading but can't locate, the citation I swore I saved. Every researcher I've talked to has a shadow library of "things I almost referenced" that vanished into a browser history.

The posts here are about closing that gap. How to clip academic papers, news sources, and primary documents as structured markdown the moment you read them. How to build a citation-ready library that survives across degree programs, jobs, or projects. How to use AI to synthesize across hundreds of saved sources without losing the trail back to the original. How to handle the messy reality of formats: arXiv HTML, JSTOR PDFs, blog post as primary source, Substack essays you want to cite.

These are practical workflows, not abstract theory about reference management. They assume you're saving as you read, not at the end of a project. They assume markdown is the lingua franca that lets the same source flow from an Obsidian vault to a Zotero library to a Claude conversation to a paper draft. The goal is one source, many surfaces. Captured once, used everywhere.