TL;DR
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
Astral Codex Ten published an October 2026 links roundup spanning AI art and translation, forecasting, marriage statistics, and other topics. The author says the links have not all been independently verified and warns that readers may identify errors or misleading claims.
Astral Codex Ten has published its October 2026 links roundup, a collection of commentary on topics ranging from AI-generated art and translation to a bot’s win in a forecasting competition and disputed claims about marriage data. The author cautions that the links have not all been independently verified and that some may be wrong or misleading, making the post a reading guide rather than a fully checked report.
The roundup opens with a historical note about Jeremy Bentham, saying he wrote a utilitarian cookbook intended for use at his Panopticon prison. The author describes its recipes as using inexpensive ingredients and notes that they include meat. Other links address a chemistry paper containing an apparently disturbing phrase, an AI artwork inspired by Hieronymus Bosch, and an attempt to render Chinese literature in English poetic meter. These are presented as links and commentary, not as independently reported investigations.
Several entries focus on AI capabilities and interpretation. The author discusses a model-assisted poetic translation of the opening of Romance of the Three Kingdoms, while acknowledging they cannot assess its fidelity to the Chinese. The post also relays a claim that an older Claude model might continue text formatted like an internal email. The author says the more sensational suggestion—that this could expose internal Anthropic emails—is probably false, and frames the output instead as a possible form of model free association.
The roundup reports that Laertes won the Metaculus Cup in September, marking the first time a bot took first place in the forecasting competition, according to the post’s description. It attributes the bot’s development to Jeffrey Liang, who reportedly spent fewer than 150 hours and a few thousand dollars on computing and data. The post also points to a new Mantic startup with $25 million in seed funding. Separately, it challenges a marriage commentary’s use of General Social Survey responses: the author says the source data show 61% of respondents describing their marriages as “very happy,” while noting that other alleged errors have not all been checked.
A Roundup With Uneven Verification
The post brings together several claims that readers may encounter separately, including AI forecasting performance and arguments about marriage and fertility. Its value is partly in directing attention to those discussions, and partly in showing where the author disputes a source’s interpretation of survey figures. The marriage example illustrates why definitions matter: classifying every answer other than “very happy” as “miserable” produces a stronger claim than the response categories alone necessarily support.
At the same time, the roundup’s breadth should not be mistaken for uniform evidence. Some items are personal reactions to creative work; others summarize outside reporting or repeat claims that remain unverified. Readers relying on the material for factual conclusions should follow the original sources and check how their figures and claims are supported.
Popular right now
KitchenAid Artisan 5-Quart Tilt-Head Stand Mixer, Almond Cream
Bose QuietComfort Wireless Noise-Cancelling Over-Ear Headphones, Black
Ninja Flip Toaster Oven and Air Fryer, 8-in-1, Stainless Steel, SP151
As an affiliate, we earn on qualifying purchases.
How the Links Are Framed
Astral Codex Ten’s post is a recurring links format: it presents multiple outside items with short reactions, rather than a single investigation with one central finding. The October edition moves across history, culture, social commentary, and technology. The source’s own disclaimer is especially relevant because its entries do not all carry the same level of verification.
For the forecasting item, the post connects a September competition result to an earlier prediction by the author about AI-human forecasting parity by the end of 2026. A bot taking first place is a discrete result; it does not by itself establish parity across forecasting tasks. The post also identifies the winning bot’s developer and reported development costs, while citing a separate article for additional information about top-performing systems.
“I haven’t independently verified each link.”
— Astral Codex Ten author, in the roundup’s verification note
Claims Still Awaiting Checks
The author explicitly says not every link has been independently verified and estimates that commenters typically find evidence that two or three links in each roundup are wrong or misleading. That estimate is the author’s account of past posts, not an audit of this edition. The author says corrections are made when identified but cannot guarantee all will be caught before readers see the post.
Several details remain unsettled in the material provided. The author says they checked the marriage survey’s “not too happy” issue but have not confirmed other alleged errors in the same commentary. They also cannot judge whether the Chinese-to-English verse preserves the original’s meaning, and the Claude email-style prompting claim is presented as a claim, not independently established evidence of model behavior or access to confidential material. The source text ends with an incomplete item, so its intended point is not clear.
Corrections and Follow-Up Links
The author says they will correct errors as they notice them and highlight important corrections later. Readers can consult the linked original articles and source material for the full evidence behind individual claims, including the Metaculus result and the General Social Survey figures. No further correction schedule or follow-up date is given in the source.
For now, the roundup should be read as a set of leads and commentary, with confidence assessed item by item. Whether the forecasting result signals broader parity, whether the disputed marriage analysis is corrected, and how the AI-related claims hold up against further checking are not resolved by the roundup itself.
Key Questions
What is the October 2026 links post?
It is an Astral Codex Ten roundup of links and short commentary on subjects including history, art, relationships, and artificial intelligence.
Did an AI bot win the Metaculus Cup?
The roundup says Laertes placed first in September, the first bot to win the competition. The post attributes the bot’s development to Jeffrey Liang; the full details should be checked against the cited competition coverage.
Does the roundup prove Claude can reveal internal emails?
No. The author describes email-style completions as a claim and says the suggestion that Claude could leak Anthropic’s internal emails is probably false. The post does not establish that the model accessed or disclosed confidential messages.
Are all the links fact-checked?
No. The author says they have not independently verified every link, and that corrections may follow when readers or the author identify errors.
What does the post say about marriage survey figures?
The author disputes a commentary’s use of General Social Survey responses, saying their reading is that 61% described their marriage as “very happy.” The author says they checked one issue but have not confirmed all alleged errors.
Source: astralcodexten.com
