SecondSourceJudgment rebuilt from primary sources

Paul Christiano, an OpenAI board member who heads ARC, a group researching how to make AI act on human intent, argues on the Dwarkesh Podcast that frontier labs should pause algorithmic progress until their security can stop the theft of weights and algorithms.

Concept illustration: on a light blue background, a dark silhouetted figure at the left holds an open hand toward a white board on two posts with a large black opening quotation mark on its top edge, and the board holds two side-by-side panels, a solid-bordered left one and a dashed-bordered right one, each with a desk and a monitor showing a speech bubble, four rising stacks of paper on a bench lettered ALGORITHMS, and a red box lettered WEIGHTS inside a large open doorway frame. In the left panel a worker in a hard hat stands on the second-tallest stack holding a sheet of paper toward the tallest one, while in the right panel no one is on the stacks, a small rail sits on top of the tallest stack, and two hard-hatted workers, one standing and one kneeling, lay a brick wall that covers the lower part of the box, beside a small pile of bricks on the floor. Printed on the image: 6-12 months Time Christiano says strengthening security to stop weight theft could take; at least a few months.

This is Christiano's personal position, and the probability attached is Christiano's subjective estimate with no model behind it. Christiano puts the chance of superintelligence within a year, on the current path without slowing down, at about one in three, and says strengthening security would take at least a few months, possibly six to twelve (Dwarkesh Podcast, 2026-11). Anthropic's amended IPO registration statement, filed November 17, also says Anthropic recently delayed pretraining of its next-generation frontier foundation model by more than a month to strengthen protection against model-weight exfiltration (Anthropic S-1/A, 2026-11-17). Host Dwarkesh Patel cited METR's result that Anthropic's rate of progress has doubled, then challenged Christiano with revenue figures: Anthropic's revenue increase in July–September was smaller than in April–June. The quarterly table in the same filing points the same way: revenue was US$4.731 billion in January–March, US$11.555 billion in April–June and US$16.956 billion in July–September, increases of US$6.82 billion and US$5.40 billion (our calculation). The quarterly table is unaudited. Christiano replied that one company's short-term revenue is "extremely noisy" and that over a longer span revenue still more than doubles every six months; on the same table, July–September revenue is 3.6 times January–March (our calculation). How we checked →

Source and checks

This item ran in the brief of Nov 17, 2026, alongside 13 other items. How we checked it, the quote checks and what was still missing that day are on that issue's How we checked page.

Cite this item

SecondSource (2026). "Paul Christiano, an OpenAI board member who heads ARC, a group researching how to make AI act on human intent, argues on the Dwarkesh Podcast that frontier labs should pause algorithmic progress until their security can stop the theft of weights and algorithms.." https://secondsource.io/en/expert/2026-11-17-1/. Accessed access date.