Emerging evidence
More
‘Ahead of his time’: AI helps unlock discoveries hidden in notebooks of man who transformed maths
- Evidence type
- News report
- Sample
- n = 663
- Sources
- 1 sourceJournalism
- Published
- 3 min read
- Truvace record
- TRV-2026-2925cite or check it
The Austrian logician Kurt Gödel was one of the most original thinkers of the 20th century. His two incompleteness theorems, published in 1931, transformed mathematics and paved the way for the digital age. After his seminal results of the 1930s, Gödel published only a few papers. His unpublished work, however, runs to tens of thousands of pages, and is of great interest to historians of science. Since 2017, the Finnish logician Jan von Plato has led a project at Helsinki University to understand the contents of…

48Gain side
How the two sides scoreClose call
46Problem side
The gain
A Helsinki University project using Claude produced a 650-page digest of Gödel's 1,009-page workbooks within days, after progress of about a page a week, and found 663 candidate theorems, of which about 400 were verified as stated in the Coq proof checker.
The problem
About 50 of the 663 theorems Claude extracted from Gödel's workbooks were false and about 200 needed repair, and the digest had to be checked page by page.
A fixed formula sets both scores from the evidence and its sources. Reader votes don’t move them.
What the scores mean
G is the gain side, out of 100. It goes up when the gain is specific, when it measurably helped people and when the sources behind it are strong.
P is the problem side, out of 100. It goes up when the problem is specific, when it measurably hurt people and when the sources behind it are strong.
In brief
Finnish logician Jan von Plato has led a Helsinki University project since 2017 to understand 16 notebooks Kurt Gödel wrote in the early 1940s, 1,009 pages of maths and logic, much of it in obsolete Gabelsberger shorthand. Progress ran at about a page a week until September, when he gave Claude his transcriptions. Within a couple of days it produced a digest of the whole material.
Asked to extract every theorem Gödel had proved, sketched or had an idea for, Claude found 663. Run through the Coq proof checker, about 400 were verified as stated, about 200 were repaired and about 50 were false. Claude also reconstructed a proof of the independence of the axiom of choice that passed Coq.
Von Plato says the workbooks hold dozens of anticipations of results not known until the 1970s and 1980s, plus 14 logic themes never pursued. Notre Dame logician Joel David Hamkins called the result remarkable if it holds up, but said historians must judge how much comes from Claude and how much from Gödel.
Main points
- The project, funded by a €2.5m European Research Council grant, covers 16 notebooks totalling 1,009 pages, much of it in Gabelsberger shorthand.
- Before Claude, the team's understanding of the text advanced about a page a week.
- Claude was given von Plato's transcriptions, six of his earlier books on Gödel and secondary literature, and returned a 650-page explanation in a couple of days.
- Of 663 theorems Claude extracted, about 400 were verified as stated, about 200 were repaired and about 50 were false, according to von Plato.
- Claude's reconstruction of Gödel's proof of the independence of the axiom of choice passed the Coq checker; von Plato says Gödel was one definition away from a correct proof.
- Von Plato says the workbooks contain dozens of anticipations of 1970s and 1980s results and 14 themes in logic not yet pursued.
- Hamkins said resemblance to Paul Cohen's proof shows Gödel was on the path, not that he had it.
The full story
Gödel's incompleteness theorems of 1931 transformed mathematics, but he published little afterwards while leaving tens of thousands of pages of unpublished work. Von Plato taught himself Gabelsberger shorthand, saying he is one of only six people in the world who can read it, and assembled a small team with a €2.5m European Research Council grant.
After meeting computer scientists at an August conference in France on AI and theorem proving, von Plato showed Claude his transcriptions. He says the AI can connect all the pieces across a thousand pages, which humans struggle to do, though every output had to be controlled page for page.
The first task was to find out whether the notebooks held a proof of the independence of the axiom of choice, which Gödel claimed in 1966 after Paul Cohen won the Fields medal. Claude's answer was 'yes, almost'. Von Plato says one definition was wrong but Gödel had the right one two years earlier, and he believes Gödel would have won the 1940 Fields medal had he published.
Von Plato plans further work, including Gödel's physics and cosmology notebooks. Hamkins has not seen the reconstruction and said that while the notebooks resemble Cohen's approach, 'you don't get credit for work you don't tell anyone about'.
What this doesn’t fix
Von Plato said Claude's output had to be controlled page for page, and logician Joel David Hamkins, who has not seen the reconstruction, said how much comes from Claude and how much from Gödel is a job for historians.
Sources
1 on file- The Guardiantheguardian.com
Talk it over
Claude produced a 650-page digest in days after progress of about a page a week, but about 50 of 663 theorems were false and about 200 needed repair. How do you weigh the speed against the checking?
‘Ahead of his time’: AI helps unlock discoveries hidden in notebooks of man who transformed mathshttps://truvace.com/item/ahead-of-his-time-ai-helps-unlock-discoveries-hidden-in-notebooks-of-man-who-tra
Claude produced a 650-page digest in days after progress of about a page a week, but about 50 of 663 theorems were false and about 200 needed repair. How do you weigh the speed against the checking?
Answer in the threadJoel David Hamkins said historians must judge how much comes from Claude and how much from Gödel. What evidence would settle that question for you?
Answer in the threadIf you were von Plato, one of only six people he says can read Gabelsberger shorthand, how would you decide which of Claude's outputs to trust without the Coq proof checker?
Answer in the thread
If you have more time
Von Plato says Gödel was one definition away from a correct proof of the independence of the axiom of choice. What would you want historians or logicians to check before accepting that claim?
Answer in the threadThe figures come from von Plato, who led the project, and Hamkins has not seen the reconstruction. Who else could review the 663 theorems, and what would you ask them?
Answer in the thread
ace
What readers say