Gödel Already Marked It Wrong

The real news

The story, straight

A mathematical proof is a chain of reasons showing why a claim follows from stated rules. A notebook can contain complete proofs, unfinished attempts and ideas that turn out wrong.

Even a genius needs a wastepaper basket.

The Guardian reports that Jan von Plato, at the University of Helsinki, gave Anthropic's Claude his transcriptions of Kurt Gödel's notebooks and supporting books. Claude did not decipher Gödel's shorthand.

A human had already done that work.

Claude produced a roughly 650-page explanation within a couple of days, which von Plato says needed checking page by page. Asked to collect claims with proofs, sketches or ideas for proofs, it listed 663 theorem candidates.

That is not 663 finished discoveries.

Von Plato says checking with Coq found about 400 held as stated, about 200 needed repairs and about 50 were false. Those figures are rounded, not an exact accounting.

Coq checks whether logical steps establish the precise statement supplied; it does not award historical credit.

The beautiful punchline: Gödel had already marked some errors himself with marginal verdicts including “falsch!” meaning “wrong”. Humanity brought computing power to a correction the pencil had already submitted.

The useful possibility is connecting ideas scattered across difficult notebooks. The limit matters too: von Plato's account is not independent confirmation, and a repaired reconstruction is not automatically Gödel's finished proof.

Study the work; keep the pencil.

Like or repost if your margin notes also deserve a research budget.

Thanks for supporting Lil'meme. Your support is appreciated.