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 roundup spanning AI forecasting, translation and art, alongside links about marriage, suburbs and historical topics. The author warns that individual links have not all been independently verified and that readers may find errors.
Astral Codex Ten has published its October 2026 links roundup, a collection of commentary on outside articles and claims that ranges from AI forecasting and translation to marriage, suburbs and historical curiosities. The post reports that an AI bot called Laertes placed first in the Metaculus Cup in September, while warning readers that the linked material has not been independently checked in every case.
The roundup’s most concrete AI development is the reported Metaculus Cup result. The author says a bot won the forecasting competition for the first time in September. Citing information reported by The Economist, the post identifies the winning system as Laertes, developed by Jeffrey Liang, who reportedly spent less than 150 hours and a few thousand dollars on computing power and data. The roundup also names systems from Mantic, FutureSearch and Preseen among the other high performers, and notes that ACX reader Ben Shindel placed first among human competitors.
The post connects that result to a broader prediction about AI-human forecasting parity by the end of 2026, which the author says they made in an earlier article. The competition result is a reported milestone, not by itself proof that AI systems have reached parity across forecasting tasks. The roundup also links to a report that Mantic raised $25 million in seed funding; the supplied material does not provide further terms or company details.
Other AI-related entries consider creative and language uses. The author praises an AI image inspired by Hieronymus Bosch, while acknowledging that its reliance on a human artist’s work complicates whether it should be treated as AI art alone. They also describe using a newer model to render the opening of the Chinese novel Romance of the Three Kingdoms in rhyming pentameter and link to an English version of the Odyssey in heroic hexameter. The author says they do not speak Chinese and presents the examples as impressions, not a formal assessment of translation quality.
AI Moves Into Forecasting and Verse
The reported competition win gives readers a specific example of AI systems being used in forecasting contests, where performance can be compared against human participants. But the roundup does not give the contest’s full scoring details, the margin between competitors or evidence that one result generalizes to other forecasting settings. The result is best understood as a notable competition outcome rather than a settled measure of broad AI capability.
The creative examples point to a different question: whether models can help produce work that meets formal constraints, such as poetic meter and rhyme, and whether that assistance expands what human artists can make. The author explicitly leaves open how much a cited image depends on its human source. Taken together, the links highlight uses worth examining while leaving judgments about quality, originality and reliability open.
Popular right now
Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation
De'Longhi Rivelia Automatic Espresso Machine with Milk Frother
Vitamix 5200 Professional-Grade Blender, 64 oz. Container, Black
As an affiliate, we earn on qualifying purchases.
A Monthly Mix of Links
The post follows a recurring links format: its author selects outside material, adds brief reactions and occasionally notes doubts or corrections. It is not a single reported investigation, and its entries differ in evidentiary status. Some describe published articles or survey data; others are personal impressions, prompts, or claims linked for readers to inspect.
That distinction matters in the entry about marriage. The roundup challenges a claim that only 18% of marriages are non-miserable, saying the cited survey appears to define “miserable” as anything short of “very happy,” including “pretty happy.” The author says their reading of the General Social Survey instead indicates that 61% described their marriage as “very happy”. They report checking the specific “not too happy” issue but say they have not confirmed other alleged errors. The post also links to arguments about American suburbs and the origins of modern polyamory, without presenting those topics as newly established findings.
Several other entries are framed as curiosities or open questions, including a historical claim about a cookbook associated with Jeremy Bentham and a riddle concerning wording in a chemistry paper. The roundup’s variety is part of its format, but readers need to distinguish the author’s commentary from the material being linked.
“I haven’t independently verified each link.”
— Astral Codex Ten author
Errors and Claims Still Open
The author says the links were not all independently verified and that commenters often identify two or three items per roundup as wrong or misleading. Corrections may be made after publication, but the author cannot guarantee every problem will have been caught by the time readers see the post.
Details of the Metaculus Cup result are limited in the supplied material: it does not give the competition’s scoring method, the field size or the gap between Laertes and other entrants. The forecast-parity prediction is the author’s stated expectation, not a confirmed outcome. The roundup also raises claims about AI-generated text, model behavior and creative quality without supplying enough material to establish broader conclusions. Its marriage-data critique is partly checked by the author, who says other errors identified by Claude remain unconfirmed.
Corrections and Forecasting Results
The author says they will correct errors as they notice them and highlight important corrections later. Readers should check the original links and any subsequent updates before relying on a disputed claim. The next meaningful test of the AI-forecasting discussion is whether further competition results and published comparisons support the suggestion of narrowing performance differences with human forecasters. The roundup itself does not announce a specific follow-up date or provide a complete evaluation plan.
Key Questions
What is the main news in the October roundup?
The roundup reports that Laertes, an AI forecasting bot, won the Metaculus Cup in September, while also collecting links on AI and other subjects. The post does not provide full competition scoring details.
Does the competition win prove AI has reached forecasting parity with people?
No. The author had predicted AI-human forecasting parity by the end of 2026, but one reported competition result does not establish parity across tasks or competitions. The roundup does not give the margin or full scoring information.
How reliable are all the links in the post?
The author says they have not independently verified every link and that readers may identify errors or misleading material. They say corrections are made as they are found, but some issues may remain.
What does the post say about the marriage statistics?
The author disputes a claim that only 18% of marriages are non-miserable, arguing that the source appears to count “pretty happy” marriages as miserable. Their reading of the General Social Survey says 61% described their marriage as “very happy”; they say they checked the specific category issue but not every alleged error.
Source: astralcodexten.com
