turning 2 years of abandoned PhD notes into an AI-searchable base instead of letting them rot
4th year NLP PhD here. I switched research directions last year and ended up with two years of paper notes, lit summaries, experiment logs, and conference notes from the old area just sitting in a ghost folder.
they weren't completely useless, but I just wasn't opening them anymore, which felt like a massive waste of all that reading and hyper-focus.
instead of letting them rot or trying to force them into a full survey paper I didn't have time to write, I tried a different route. I dumped my old transformer notes and summaries into a public library using Linkly AI, basically making the whole folder something people can actually query.
it actually paid off last week. A collaborator asked about some background context in that old area. Instead of sending them a messy 100MB zip file or a list of 20 papers they'd never read, I just shared the library link and told them to run their questions through Claude against it. The answers actually surfaced directly from my old reading insights.
idk if this is a standard workflow for anyone else, but it feels like a solid middle ground for academic hoarders. Anyone else doing something similar with their abandoned literature graveyards?