A paper, taken apart

Every paper is read as a claim-evidence graph — each finding paired with how they know it: the method, the sample, the effect size, the uncertainty. A weak paper gets an honest teardown, not a hype summary.

No papers on the desk yet.