Latest Corrections
X Sep 9, 2026 at 08:20 PM
x.com/XMihura/status/2097231283394867417
- Eso exagera el estado real del hallazgo: OpenAI anunció una solución propuesta, pero el problema de Navier‑Stokes no está oficialmente confirmado como resuelto.
X Sep 9, 2026 at 08:16 PM
x.com/OpenAI/status/2095595742975197690
- OpenAI's own launch benchmarks and independent benchmarking did not show GPT-6 Astra leading all of those domains. In the launch table, Astra trailed competing models on several software-engineering and science benchmarks, so the blanket 'state of the art' claim across every listed area is overstated.
X Sep 9, 2026 at 08:15 PM
x.com/FinanzasNoobie/status/2096957446140940565
- La cifra de 1.433.847,58 € no corresponde solo a la web pública. Según la documentación oficial, ese importe es el contrato completo de un sistema integral de gestión durante 4 años, que además de un portal ciudadano incluye consultoría, CRM y mantenimiento.
X Sep 9, 2026 at 08:11 PM
x.com/bryan_caplan/status/2085022634551451778
- Standard international measures do not show Denmark or Norway having very low living standards relative to the U.S. UNDP ranks Norway and Denmark above the U.S. on the 2023 Human Development Index, and its income table places both countries at or above the U.S. on estimated GNI per capita.
Wikipedia Sep 9, 2026 at 08:04 PM
en.wikipedia.org/wiki/Class_of_1999
- The film does not identify the program as a CIA experiment. Contemporary film records and transcript text say the robot-teacher program is sponsored by the Department of Educational Defense and MegaTech.
Substack Sep 9, 2026 at 07:55 PM
www.betonit.ai/p/tabarrok_shouldhtml
- This is too broad: not all universities are nonprofits. U.S. higher education includes public universities, private nonprofit universities, and private for-profit universities.
Wikipedia Sep 9, 2026 at 07:37 PM
en.wikipedia.org/wiki/Charlie_and_the_Chocolate_Factory
- The UK edition was not published in November 1964. Multiple bibliographic sources place the first UK publication with George Allen & Unwin in 1967.
- The UK publication did not follow 11 months after the 1964 US release. The first UK publication was in 1967.
- George Allen & Unwin was not the original publisher of the book. The original 1964 edition was published by Alfred A. Knopf in New York; George Allen & Unwin published the later UK edition.
Wikipedia Sep 9, 2026 at 07:37 PM
en.wikipedia.org/wiki/Willy_Wonka_&_the_Chocolate_Factory
- The Oscar nomination was not for Best Original Score. The film was nominated in the separate category for Scoring Adaptation and Original Song Score.
- This plot summary assigns the final line to Charlie, but it is Wonka who tells Charlie that line at the end of the film.
X Sep 9, 2026 at 07:05 PM
x.com/Dr_Singularity/status/2097746969212981732
- This overstates the current status. As of September 9, 2026, the Clay Mathematics Institute still lists the Navier-Stokes Millennium problem as unsolved, and its rules require publication, a two-year wait, and general acceptance before a proposed solution can be considered.
X Sep 9, 2026 at 06:46 PM
x.com/BetterCallMedhi/status/2097472336257863722
- This misstates the official Clay problem. The Clay Millennium Prize formulation explicitly allows a smooth external force in its breakdown alternatives, and Euler is not itself one of Clay’s Millennium Prize problems.
X Sep 9, 2026 at 05:41 PM
x.com/esaagar/status/2029979781576413468
- The 400,000 figure understates Soviet 1945 military casualties. Archival-based military histories put the Red Army at 194,191 casualties in the Vistula–Oder offensive alone and more than 361,000 in the Berlin operation alone.
Wikipedia Sep 9, 2026 at 05:01 PM
en.wikipedia.org/wiki/Dry_lab
- This misdefines proteomics. Proteomics is the large-scale study of all proteins in a cell, tissue, or organism, not specifically the elucidation of unknown protein structures and folding patterns.
- This is misleading because Ramachandran plotting is not a protein-structure determination technique. A Ramachandran plot is mainly a tool for visualizing allowed backbone angles and for validating already solved or modeled protein structures.
Substack Sep 9, 2026 at 08:16 AM
bafriedman.substack.com/p/recon-strike-battle
- The official Marine Corps concept is called Expeditionary Advanced Base Operations (EABO), not "Expeditionary Advance Base Operations."
Wikipedia Sep 9, 2026 at 07:51 AM
en.wikipedia.org/wiki/If_Anyone_Builds_It,_Everyone_Dies
- The page count is wrong. Official publisher listings for this book’s hardback edition give 272 pages, not 256.
Substack Sep 9, 2026 at 07:17 AM
thezvi.substack.com/p/frontier-lab-employee-open-letter?r=67...
- Anthropic’s Claude Code creator is Boris Cherny, not “Boris Cherney.”
Substack Sep 9, 2026 at 07:12 AM
thezvi.substack.com/p/the-pacing-of-the-frontier
- OpenAI did not say it had solved 10 open math problems; its August 1 announcement said the ten results either resolved or made substantial progress on long-standing problems, including theoretical computer science problems.
LessWrong Sep 9, 2026 at 06:56 AM
www.lesswrong.com/posts/6j3kBHdowGLCeqobg/dear-god-please-do...
- Kokotajlo later said he ultimately kept the equity, and OpenAI said it never canceled vested equity. He was at risk of losing the money, but did not in fact lose it.
X Sep 9, 2026 at 05:55 AM
x.com/lu_sichu/status/2097444147485720580
- This is incorrect: Tao did post about the proof itself, summarizing the result and its main ideas on Mastodon and in a blog post.
X Sep 9, 2026 at 04:21 AM
x.com/allTheYud/status/2097405619720573334
- This misdescribes the paper. The referenced work was not an 18-month longitudinal study of Llama 3.2; Llama 3.2 was released on September 25, 2024, and the paper was first submitted on January 24, 2025, while the paper’s “18-month” wording refers to benchmark data spanning historical months.
Substack Sep 9, 2026 at 04:11 AM
partialagonism.substack.com/p/if-ai-is-so-smart-why-hasnt-it...
- This is incorrect because daptomycin (Cubicin), approved later in 2003, was also described by Merck as a blockbuster antibiotic with about $1 billion in trailing-12-month U.S. revenue.