What's this about?
A sign error invalidated an argument and two dependent results. The repository also records repairs to 14 manuscripts; 300 of 719 top-line results are formalized, so computer checking does not cover the whole collection.
What exactly is formal proof?
A proof written in a language whose logical steps a computer can check. Correctly stating the theorem and its assumptions still matters.
What's the community saying?
HN 256 points, 550 comments
- Which withdrawn results had actually been Lean-checked, versus merely written as mathematical prose?
- Can a formally checked statement still miss the intended theorem through wrong assumptions?
- Does rapid publication shift too much verification work onto human mathematicians?