Skip to content
Theorem Weekly TW markTWTheoremWeekly

The evidence behind the week in machine learning.

Model updates

Can LLMs Decode Unpublished Logic Notebooks Faster Than Human Scholars?

AI model Claude summarized 1,009 pages of Kurt Gödel’s shorthand notebooks in days, revealing hundreds of previously unknown proved theorems.

Key points

  • Jan von Plato used Claude to digest 1,009 pages of Gödel’s 1940s workbooks in September 2026.
  • The AI output revealed hundreds of proved theorems and 14 unpursued logic themes.
  • Human progress had averaged one page per week prior to the AI intervention.
Can LLMs Decode Unpublished Logic Notebooks Faster Than Human Scholars?
Depth
1 outlet · newest source 66h ago

Jan von Plato reported using the Claude model to analyze Kurt Gödel’s unpublished notebooks in September 2026, according to a Guardian article. The Finnish logician had led a project since 2017 to decipher 16 notebooks written by Gödel in the early 1940s.An AI tool helped a researcher understand thousands of pages of old math notes much faster than before.Large language models demonstrated high-fidelity extraction of mathematical proofs from obsolete shorthand notation.

Evidence: reportedAI unlocks Gödel's notebooks, said to contain hundreds of proved theorems

aifleet-hn-ai

Evidence: reportedAI unlocks Gödel's notebooks, said to contain hundreds of proved theorems

aifleet-hn-ai

What was released or claimed

Von Plato stated that Claude provided a comprehensive digest of the entire 1,009-page collection within a couple of days. This rapid processing revealed that Gödel had proved hundreds of theorems previously unknown to the scholarly community.

Who it affects

Historians of science and logicians studying Gödel’s unpublished work are the primary beneficiaries. The findings impact researchers interested in the development of mathematical logic and the history of computation theory during the mid-20th century.

What is still unknown

The specific content of the 14 themes in logic that have not been pursued remains undefined in the report. The full verification of the hundreds of proved theorems by independent human scholars has not been detailed in the available sources.

So what

The speed difference between human and AI analysis of dense, obsolete shorthand is stark. Von Plato described previous progress as painfully slow, averaging about one page per week before the AI intervention in September 2026.

What the paper shows

The Guardian report indicates that Claude successfully interpreted Gabelsberger shorthand, a German system obsolete even when Gödel used it. The model identified mathematical structures and proofs that human scholars had missed or could not process at scale. Von Plato expressed surprise, noting Gödel was far ahead of his time with these discoveries.

Why it matters

  • Demonstrates the feasibility of using LLMs for deciphering obsolete technical shorthand in historical archives.
  • Highlights the potential for AI to uncover latent scientific knowledge in dense, unpublished manuscripts.
  • Provides a case study on the acceleration of humanities research through machine reading comprehension.
  • Raises questions about the verification standards for AI-discovered mathematical theorems.

Topics: large language models, historical archives, mathematical logic, AI applications

Updated

Then / now / next

Then

Human scholars were deciphering Gödel’s notebooks at a rate of about one page per week.

Now

AI model Claude has summarized the entire collection, revealing hundreds of unknown theorems.

Next

Researchers will need to verify the AI-discovered proofs and explore the 14 unpursued logic themes.

Story timeline

  1. Then Human scholars were deciphering Gödel’s notebooks at a rate of about one page per week.
  2. 13 Nov 2026 AI unlocks Gödel's notebooks, said to contain hundreds of proved theorems
  3. Now AI model Claude has summarized the entire collection, revealing hundreds of unknown theorems.
  4. Next Researchers will need to verify the AI-discovered proofs and explore the 14 unpursued logic themes.

Ask the article

Local keyword search of this article; no generated answers.

Did this article give you the evidence you needed?

No vote saved on this device.
Your choice stays on this device. No public tally.

Questions readers ask

How long did it take Claude to process the notebooks?

Claude provided a comprehensive digest of the entire 1,009-page material within a couple of days.

What shorthand system did Gödel use?

Gödel wrote much of the text in Gabelsberger, a German shorthand system that was obsolete during his lifetime.

How many new theorems were found?

The AI analysis revealed that Gödel had proved hundreds of theorems that no one knew about previously.

Connected stories

This storyCan ERNIE 5.5 Flash Base match DeepSeek V4 on open benchmarks?Can ERNIE 5.5 Flash Base match DDid AI enable a 96-year-old’s solo Tibet trip?Did AI enable a 96-year-old’s soDoes AI-Generated Code Create Fragile SaaS Systems?Does AI-Generated Code Create FrForum AI study finds mainstream chatbots misstate election facts nearly 40 percent of the timeForum AI study finds mainstream

Reading receipt

  • Jan von Plato used Claude to digest 1,009 pages of Gödel’s 1940s workbooks in September 2026.
  • The AI output revealed hundreds of proved theorems and 14 unpursued logic themes.
  • Human progress had averaged one page per week prior to the AI intervention.

Still open: How will the mathematical community verify the AI-identified theorems?

Still open: What are the specific 14 logic themes Gödel explored but did not publish?

Time spent: 0:00 · Depth: Full

Sources

  1. Hacker News: AI posts over 50 points

Related stories

Background reading