Jack Clark · Sep 21, 2026
Jack Clark’s Import AI surveys AI strategy, research, and possible futures
Markdown · JSON · Original source
Jack Clark published this Import AI newsletter on September 21, 2026. He reports that a RAND paper recommends a US “Freedom of Action” strategy: spend now to preserve options amid uncertainty about the path to superintelligence. Clark contrasts that approach with what he sees as a US strategy focused on acceleration, comparing it to improving a car’s engine without investing in safety measures. The newsletter also covers research on human brain tissue transplanted into mice, a proposed research agenda for pacing AI progress, a study of uncensored open-weight models, and Toby Ord’s analysis of recursive self-improvement (RSI). Ord argues that resource and time constraints could limit how quickly AI systems improve, while warning that even faster progress could bring dangers. Clark’s final Tech Tales is explicitly speculative: it imagines future researchers trying to connect machine-made science with human knowledge. Together, the pieces raise questions about how governments might prepare for uncertain AI progress and how researchers might study its possible effects. The newsletter presents competing arguments about pacing, but does not establish which policy choices will work. What evidence would help policymakers assess whether a pacing intervention is reducing risks without imposing greater costs?
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Original source
Jack Clark · Article
Source details
Original source date: Sep 21, 2026 (2026-09-21T12:31:04.000Z).
Date basis: Date supplied by the publisher’s feed.
Added to Escape Meridian: 2026-10-29T16:58:19.244Z. First observed: 2026-10-23T10:35:58.231Z. Retrieved: 2026-10-23T10:36:12.847Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
Text as retrieved on October 23, 2026; originally published September 21, 2026. The page is marked edited after publication. Later additions may be present; claims here are limited to material supported by the retrieved text and are not presented as later developments. Clark’s policy comparisons and the Tech Tales passage are his commentary and fiction, respectively, not demonstrations of policy outcomes or future events. The newsletter summarizes several underlying papers; this reading does not independently assess their methods or findings. Text as retrieved on October 23, 2026; originally published September 21, 2026.
Open question
What evidence would help policymakers assess whether a pacing intervention is reducing risks without imposing greater costs?
Source-stated
Clark reports that RAND recommends a US “Freedom of Action” strategy to preserve options amid uncertainty about the path to superintelligence.
because so much of how the next phase of the AI takeoff will occur is unknown, the best thing is for the US to maintain a “Freedom of Action” strategy.
Source-stated
Clark characterizes the current US approach as focused on acceleration and argues that preparation should include spending on safety measures.
Right now, it feels like the US strategy can mostly be described as the “acceleration” one described here by RAND
Source-stated
The researchers behind a pacing agenda describe reasons both for and against pacing AI progress, including risk reduction and delayed benefits.
Arguments against pacing: Delays benefits like healthcare and other science-driven breakthroughs; can potentially increase concentration of power
Source-stated
Toby Ord argues that resource and time constraints could limit the speed of recursive self-improvement, while noting that faster progress could still create dangers.
There are many kinds of feedback loop that could contribute to RSI, ranging from decades (e.g., designing a successor to EUV lithography) to months (e.g., designing better pretraining) to seconds (e.g., designing better scaffolds).
Escape Meridian interpretation
Escape Meridian reads the newsletter’s policy and research coverage as raising a shared question: how to prepare for uncertain AI progress without assuming that one policy approach will work.
The five main uncertainties: Danger proximity: At what point does AI become so dangerous the risks aren’t worth the benefits?
Source and contributor context
Jack Clark · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media retrieval status: unchecked.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Superintelligence
A hypothetical AI far more capable than people across most intellectual tasks.
Open weights
A model whose trained numbers are published so others can download, run and inspect it; the license and training data may still be restricted.
Jack Clark
Jack Clark writes Import AI, a newsletter covering AI research, development and its implications.
Editorial provenance
Version 1. Publication ID: 4718507a-0746-8301-3612-9b8059cb5e69.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: passed by anthropic/claude-opus-5.5 on 2026-10-29T16:56:20.787Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_screen.
Observation: ecbe9a354ce60b8a451f08c48affb33bdc85925b906d139ea78b3bec628eb2d6. Source text hash: e1716b4c54a294dd5767fb8582f3bfa7ddc7a2c4bf516c2fb992161538b4b856.
Revision and source-check history
Version 1 · 2026-10-29T16:58:19.244Z
Jack Clark’s Import AI surveys AI strategy, research, and possible futures
Jack Clark published this Import AI newsletter on September 21, 2026. He reports that a RAND paper recommends a US “Freedom of Action” strategy: spend now to preserve options amid uncertainty about the path to superintelligence. Clark contrasts that approach with what he sees as a US strategy focused on acceleration, comparing it to improving a car’s engine without investing in safety measures. The newsletter also covers research on human brain tissue transplanted into mice, a proposed research agenda for pacing AI progress, a study of uncensored open-weight models, and Toby Ord’s analysis of recursive self-improvement (RSI). Ord argues that resource and time constraints could limit how quickly AI systems improve, while warning that even faster progress could bring dangers. Clark’s final Tech Tales is explicitly speculative: it imagines future researchers trying to connect machine-made science with human knowledge. Together, the pieces raise questions about how governments might prepare for uncertain AI progress and how researchers might study its possible effects. The newsletter presents competing arguments about pacing, but does not establish which policy choices will work. What evidence would help policymakers assess whether a pacing intervention is reducing risks without imposing greater costs?
Original source
Jack Clark · Article
Source details
Original source date: Sep 21, 2026 (2026-09-21T12:31:04.000Z).
Date basis: Date supplied by the publisher’s feed.
Added to Escape Meridian: 2026-10-29T16:58:19.244Z. First observed: 2026-10-23T10:35:58.231Z. Retrieved: 2026-10-23T10:36:12.847Z.
Limits
Text as retrieved on October 23, 2026; originally published September 21, 2026. The page is marked edited after publication. Later additions may be present; claims here are limited to material supported by the retrieved text and are not presented as later developments. Clark’s policy comparisons and the Tech Tales passage are his commentary and fiction, respectively, not demonstrations of policy outcomes or future events. The newsletter summarizes several underlying papers; this reading does not independently assess their methods or findings. Text as retrieved on October 23, 2026; originally published September 21, 2026.
Open question
What evidence would help policymakers assess whether a pacing intervention is reducing risks without imposing greater costs?
Source-stated
Clark reports that RAND recommends a US “Freedom of Action” strategy to preserve options amid uncertainty about the path to superintelligence.
because so much of how the next phase of the AI takeoff will occur is unknown, the best thing is for the US to maintain a “Freedom of Action” strategy.
Source-stated
Clark characterizes the current US approach as focused on acceleration and argues that preparation should include spending on safety measures.
Right now, it feels like the US strategy can mostly be described as the “acceleration” one described here by RAND
Source-stated
The researchers behind a pacing agenda describe reasons both for and against pacing AI progress, including risk reduction and delayed benefits.
Arguments against pacing: Delays benefits like healthcare and other science-driven breakthroughs; can potentially increase concentration of power
Source-stated
Toby Ord argues that resource and time constraints could limit the speed of recursive self-improvement, while noting that faster progress could still create dangers.
There are many kinds of feedback loop that could contribute to RSI, ranging from decades (e.g., designing a successor to EUV lithography) to months (e.g., designing better pretraining) to seconds (e.g., designing better scaffolds).
Escape Meridian interpretation
Escape Meridian reads the newsletter’s policy and research coverage as raising a shared question: how to prepare for uncertain AI progress without assuming that one policy approach will work.
The five main uncertainties: Danger proximity: At what point does AI become so dangerous the risks aren’t worth the benefits?
Source and contributor context
Jack Clark · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media retrieval status: unchecked.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Superintelligence
A hypothetical AI far more capable than people across most intellectual tasks.
Open weights
A model whose trained numbers are published so others can download, run and inspect it; the license and training data may still be restricted.
Jack Clark
Jack Clark writes Import AI, a newsletter covering AI research, development and its implications.
Editorial provenance
Version 1. Publication ID: 4718507a-0746-8301-3612-9b8059cb5e69.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: passed by anthropic/claude-opus-5.5 on 2026-10-29T16:56:20.787Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_screen.
Observation: ecbe9a354ce60b8a451f08c48affb33bdc85925b906d139ea78b3bec628eb2d6. Source text hash: e1716b4c54a294dd5767fb8582f3bfa7ddc7a2c4bf516c2fb992161538b4b856.
passed · 2026-10-29T16:56:20.787Z · anthropic/claude-opus-5.5
Publication: 4718507a-0746-8301-3612-9b8059cb5e69
attribution — pass: The reading attributes the newsletter to Jack Clark, the RAND strategy to a RAND paper as reported by Clark, the pacing agenda to its researchers, and the RSI analysis to Toby Ord. The Tech Tales is correctly identified as Clark's fiction. All attributions match the source.
RAND has published a lengthy paper about what the US should do to “secure geopolitical advantage on an uncertain path to superintelligence”.
summary — pass: The summary leads with the actor (Clark) and the main item, RAND's Freedom of Action strategy, explained in plain terms. It also covers Clark's acceleration critique and the car analogy. Secondary items are named briefly, and RSI is spelled out and glossed as how quickly AI systems improve. The title is generic but accurate. The question follows from the pacing material and is not leading. A non-specialist can follow the central meaning.
which is analogous to me to sitting in a car and spending all your resources on making the car go faster and upgrading the engine, and nothing on proactive safety measures
date — pass: The reading states September 21, 2026, which matches the UTC day of originalPublished (2026-09-21T12:31:04Z). The kind is publisher-page, so no archive disclosure is needed. The page was edited after publication, but the cited material is the newsletter's core body and plainly part of the original piece. The future dates in Tech Tales (2032) are fiction and are labeled as speculative, not presented as events. There is no hindsight framing.
support — pass: Each claim is grounded in the source text. RAND's recommendation, Clark's acceleration characterization and his call for spending, and the arguments for and against pacing are all stated. Ord's limits thesis and his warning about danger also appear in the text, though the excerpt chosen for that claim speaks only to feedback-loop timescales. The interpretation is labeled as Escape Meridian's own, and it avoids presenting any policy as proven. No benefits or costs are invented.
that doesn’t mean RSI is safe or that AI R&D will move at a manageable pace
media — pass: No media was selected and no verified candidates were supplied, so there is no identity or relevance problem.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.