Corrections

Last updated October 31, 2026

When something on this site turns out to be wrong, it is fixed and listed here with the date and the source used. To report an error, email charles@digitalsuperintelligence.us.

October 31, 2026

• How many zeros have been checked. The Riemann dossier said the first 10 trillion zeros had been checked; the current figure is 12 trillion. (Quanta)
• When Hodge asked. The Hodge tile said "Since 1950", the year Hodge presented the conjecture at the International Congress; the dossier dates it to 1941, when he first posed it. The tile now says 1941. (Clay)
• The Navier–Stokes door. The page said the first door moved on September 30 (the quasi-Riemann proof) while its own banner reported Clay’s September 11 statement on Navier–Stokes, and door 03 had no news after 2025. The intro and door 03 now give OpenAI’s September 8, 2026 claimed blowup proof, Clay’s review and the credit dispute. (Clay)
• Sunlight on Earth. The Kardashev paragraph said about 10¹⁶ W of sunlight reaches Earth. That figure is Sagan’s Type I benchmark; the sunlight striking Earth is about 1.7×10¹⁷ W. (Universe Today)
• Kardashev and CTA-102. His suggestion that CTA-102 might carry signals was dated 1963; the paper is 1964. (SETI Institute)
• The 2026 matter-wave record. The new record was given as 170,000 proton masses; the source gives more than 170,000 atomic mass units. (Physics World)
• Who made the quasi-Riemann proof. The Riemann door called it an internal OpenAI model; the source says only an OpenAI model. (Quanta)
• The quasi-Riemann proof. The page said the first AI proof came in October 2026 and the shorter version followed on Oct 5. The first version is dated September 30, 2026; the shorter version (zeros between 1/12 and 11/12) followed about a week later and was released Oct 6. (Quanta)
• The matter-wave record. The double-slit timeline called the 2019 Vienna result (molecules over 25,000 proton masses) the record. In 2026 the same group showed interference for sodium clusters of more than 170,000 atomic mass units. (Physics World)
• A Hodge preprint. A news item about an August 2025 preprint giving a conditional reduction for threefolds was removed: the rational Hodge conjecture for threefolds is already a theorem, so the item was not news. (Clay)
• Where Hou spoke. Thomas Hou's April 2, 2025 talk was credited to Berkeley Lab. It was the UC Berkeley Applied Math Seminar. (seminar page)
• Atiyah and Hirzebruch. Codex Obscura dated their counterexample to the integral Hodge conjecture 1961. The paper, Analytic cycles on complex manifolds, appeared in Topology in 1962. (Topology)
• The Kardashev numbers. The paragraph gave humanity about 0.73 and Type I as about 4×10¹² W without saying these come from two different scales. It now says the 4×10¹² W figure is Kardashev's original 1964 Type I and that Sagan's continuous version sets Type I at 10¹⁶ W. (Universe Today)
• Who proved quantum gravity non-renormalizable. The page credited Feynman and DeWitt. The results are 't Hooft and Veltman (1974) and Goroff and Sagnotti (1986). (Nuclear Physics B)
• What Karp showed. The page said Karp showed almost every NP problem without a fast algorithm is NP-complete. In 1972 he proved 21 well-known problems NP-complete. (Karp, 1972)
• The age of the universe. The worked answer now reads 13.80 billion years (Planck 2018 gives 13.787 ± 0.020). (Planck 2018)
• Where arXiv began. A resource note said arXiv was founded by Cornell. Paul Ginsparg founded it at Los Alamos in 1991; Cornell has run it since 2001. (arXiv)