most of us dive into research papers the wrong way: start at page 1, get lost by page 3, and give up by page 5.
keshav’s 2007 piece flips that around with a three-pass method:
- first pass (5–10 min): scan the title, abstract, intro, headings, conclusion, and references. just enough to know if the paper is worth your time.
- second pass (up to 1 hour): read with care, look at figures and results, jot notes, but skip proofs and heavy math. now you should be able to explain the gist to someone else.
- third pass (deep dive): re-create the work in your head, check assumptions, re-trace logic, and challenge the results. this is where you really “own” the paper.
he also shows how this scales to doing a literature survey: start broad, skim lots of papers with one pass, and narrow down only when needed.
the key takeaway: not every paper deserves the same level of attention. adjust your depth depending on your goals, and you’ll save hours while actually learning more.
sometimes the smartest move isn’t reading harder, it’s reading smarter.
full paper: how to read a paper (PDF)