News for hackers. Stuff that matters.
Re:Claude Science by cma on Tuesday June 30, 2026 @01:31PM (#48736163)
In some fields like comp sci, when code isn't given but the paper describes the approach, LLMs do help with the reproducibility crisis: you can ask it to reproduce the result through reimplementation by reading the paper.

If it fails you may have to double check it did properly reimplement it, but if it succeeds you do get a reproduction.

There are no comments on this comment yet.