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 a November 2026 links roundup covering AI risk forecasts, California flood predictions, debates about a future without work and other topics. The author cautions that links have not all been independently verified and that readers may identify errors or misleading summaries.
Astral Codex Ten published its November 2026 links roundup, a collection of outside articles, forecasts and observations that includes claims about AI-related risks and California flood prospects. The author says the links have not all been independently verified and warns that readers commonly find two or three inaccurate or misleading items in each roundup.
The entry receiving the most extensive treatment concerns forecasts from Sentinel, described by the author as an ACX grantee that forecasts global disasters. The post reports Sentinel’s estimate of a 25% chance that a rogue AI will exfiltrate its weights over the next 12 months. That is a forecast, not evidence that such an event has occurred, and the post does not provide details about the forecasting method or how the event is defined.
The roundup also relays a warning from a Sentinel-affiliated forecaster that California is expected to have a much rainier-than-normal winter. The forecaster is quoted as assigning a 2.5% chance to a “major flood.” The author then describes asking an AI forecaster to translate that broad estimate into possible damage scenarios: an 8% chance of flooding causing at least $10,000 in damage to at least 100,000 homes, and a 2% chance of that level of damage affecting at least one million homes. Those figures are the author’s account of an AI-generated interpretation, not a directly stated Sentinel forecast.
The author further says the AI forecaster put the chance of an ARkStorm this year at 5%, compared with a 1% baseline, and that roughly half of those scenarios might reach the one-million-home damage threshold. The post does not establish that these estimates are independently validated. Other entries range from a discussion of the “Wow signal” in radio astronomy to an essay about how people might spend time in a world with less productive work, and a reported meeting between two Satmar Rebbes.
Forecasts Raise Preparedness Questions
The California entry matters because even a low-probability disaster forecast can inform household preparation, while the post’s layered estimates make clear how much interpretation can intervene between an initial warning and a specific damage scenario. The author urges readers in California to check roofs and gutters ahead of a wetter winter. The reported probabilities should not be treated as an official government warning or as a settled estimate of losses: the source does not name a public agency issuing an alert or provide the underlying forecasting documentation.
The AI forecast also draws attention to a separate question: how to interpret a numerical prediction about a rare event when its definition and method are not explained in the post. A 25% estimate for AI weights being exfiltrated is not a report of a breach, and the roundup supplies no evidence that such an event is underway. Keeping those distinctions clear helps readers understand the entries as forecasts and commentary rather than confirmed developments.
Across the collection, the post is best read as a guide to material the author found interesting, not as a single investigation or independently checked news report. Its own verification warning is especially relevant where the roundup paraphrases external writing or presents numbers without linked methodology in the supplied text.
Popular right now
Ninja DualBrew Pro Specialty Coffee Maker, Black, CFP301
KitchenAid Artisan 5-Quart Tilt-Head Stand Mixer, Almond Cream
Philips Café Aromis Automatic Espresso Machine with LatteGo Pro Milk System
As an affiliate, we earn on qualifying purchases.
How the Links Post Is Framed
The November post follows the format of an ongoing links roundup: it points readers to outside reporting, essays, social-media observations and forecasts, then adds the author’s reactions. The entries are not presented as one unified argument. Topics include nominative determinism, Soviet-era acronym names, anagrams of public figures’ names, the lab-leak debate, UFO skepticism and the Berkeley rationalist and effective-altruism communities.
Some entries summarize a source; others add the author’s own interpretation or an informal follow-up. In the California section, for example, the rain outlook and “major flood” probability are attributed to a forecaster, while the more detailed household-damage estimates are described as an AI forecaster’s operationalization. The post’s author also states that readers typically identify errors or misleading claims after publication, and says corrections are made when found. That caveat is part of the source’s framing, not proof that any particular link is wrong.
The material supplied for this article ends during an entry about a critique of natural-law arguments. It does not include the full roundup or establish whether additional entries followed. No independent source material accompanies the post, so its summaries and forecasts cannot be verified here against the original linked items.
“I haven’t independently verified each link.”
— Astral Codex Ten author, in the roundup’s opening disclaimer
Forecast Methods Remain Unspecified
The roundup does not provide the methodology, source documents or calibration record behind Sentinel’s AI exfiltration estimate or the California flood probabilities. It also does not identify exactly what qualifies as a “major flood” in the 2.5% forecast. The damage scenarios are explicitly described as an AI forecaster’s interpretation, and the source gives no basis for assessing how that system produced them.
It is not clear from the supplied material whether the winter outlook or forecasts have since changed, whether an official agency has issued a related warning, or whether the predicted conditions have occurred. The author’s opening caveat also means the summaries of linked articles should not be treated as independently confirmed facts. The source excerpt ends partway through its final listed item, so the full scope of the roundup is unavailable here.
Corrections and Weather Updates
The author says they correct errors as they are identified and plan to highlight important corrections later. Readers should check the original linked forecasts and subsequent updates before relying on the roundup’s numerical estimates. For California weather risks, the next useful developments would be updated forecasts from relevant weather and emergency-management agencies, along with clearer definitions of the flood scenarios cited.
The AI-risk estimate remains a forecast covering the next 12 months, not a scheduled event or confirmed incident. The post gives no further milestone for that prediction. Any assessment of whether it remains current would require updated information from Sentinel or other sources.
Key Questions
What is the November 2026 links post?
It is an Astral Codex Ten roundup of outside links and commentary on subjects including AI forecasts, flood risk, science, work and community life.
Did Sentinel report that a rogue AI had stolen its weights?
No. The roundup reports a 25% forecast for the next 12 months that a rogue AI might exfiltrate its weights. It does not describe a confirmed incident.
What does the post say about California flooding?
It relays a forecaster’s estimate of a 2.5% chance of a major flood amid expectations of a much rainier-than-normal winter. More specific home-damage probabilities are described as an AI forecaster’s interpretation, not the original forecast.
Are the roundup’s claims independently verified?
No. The author explicitly says they have not independently verified every link and warns that readers often identify errors or misleading summaries. The post says corrections are made when found.
Source: astralcodexten.com
