What Just Happened With the Pacing AI Debate (TBPN)

Author: Brandon Gorrell and Jackson Fordyce (TBPN) URL: https://open.substack.com/pub/tbpn/p/what-just-happened-with-the-pacing One-line: Over one weekend the AI doom debate became an AI slowdown debate: Dario Amodei posted “We Must Pace the Frontier,” the other lab heads agreed within hours, and the critics (Sacks, Baker, Thompson) asked why the frontier labs need anyone’s permission to slow down, and whether pacing can work at all with China in the race.

Key claims

• Dario’s plan has three steps. (1) Give independent third-party evaluators access to AI companies so they can verify safety practices. (2) Frontier labs in democratic countries work together, with government help, to set AI safety standards. (3) The U.S. and other democracies try to coordinate with authoritarian governments.
• The consensus formed within hours. Timeline as TBPN gives it, all PT: Saturday 7am, Dario’s blog. 8am, Elon Musk agrees. 9:30am, Sam Altman agrees. 3:59pm, Demis Hassabis broadly agrees but says it needs work. 8:14pm, David Sacks raises questions. Sunday 9:11am, the Financial Times reports Donald Trump rejects tech bosses’ calls for a slowdown. 9:17am, Gavin Baker’s round-up. Monday morning, Ben Thompson’s Stratechery response.
• “The timeline” pushed back with practical questions. Who will the third-party evaluators be? Why wait for Washington to regulate the frontier? If Anthropic believes AI is advancing too quickly, why not simply slow down on its own?
• Sacks: you are the frontier, so just slow down. Sacks says Anthropic and OpenAI already hold a duopoly on frontier intelligence by market share, revenue growth and model capability. He supports their slowing down, but rejects any need for others’ permission, an antitrust carve-out to form a “cartel”, a regulatory approval process that supersedes product liability, or evaluators policing non-frontier competitors. He says METR is “intertwined with Anthropic’s investors and staff.”
• Baker: expect an executive order, and avoid centralization. Baker’s round-up collects what Dario, Sam, Elon, Alexandr Wang, Clem (Clem Delangue) and others said. He predicts an executive order whose language will matter, and argues AI can be distributed broadly and safely rather than centralized in a few corporations that “might each become more powerful than any single government.”
• Thompson: pacing is futile without China, and political. Thompson argues any slowdown falls apart if China won’t slow too, which makes oversight of frontier labs ineffective and potentially partisan. He calls it an attempt by “a specific political movement in San Francisco, moving in conjunction with one of the major political parties” to insert itself into every AI company.
• The surrounding headlines show how far the story spread. The issue’s links include China’s state newspaper calling Anthropic’s slowdown push a “Cold War” tactic (Reuters), the WSJ’s “China Also Thinks AI Could Kill Us. But First, It Wants to Match the U.S.”, Nvidia weighing $10B of backing for an Anthropic IPO (Bloomberg/Reuters), former FTC commissioner Lina Khan on AI regulatory capture, Tyler Cowen on “The mathematicians rebel against AI”, and the FT on JPMorgan cutting off lending to Situational Awareness after AI losses.

Notable quotes

“Over the weekend, the AI doom debate turned into the AI slowdown debate with Dario sharing a blog Saturday morning titled ‘We Must Pace the Frontier.’”

“You guys are the frontier. By any reasonable metric — market share, revenue growth, model capability — the two of you have a duopoly on frontier intelligence. You’ve also claimed the lead is widening because of recursive self-improvement.” — David Sacks

“But stop pretending you need anyone else’s permission. Stop pretending antitrust law has to be suspended so you can form a cartel. Stop pretending you need a regulatory approval process that supersedes product liability. Stop pretending METR is independent when it is intertwined with Anthropic’s investors and staff.” — David Sacks

“An executive order seems likely after all this and the language in this EO is going to be really important. It is possible to democratize and distribute AI broadly and safely without centralizing it in the hands of a few corporations who might each become more powerful than any single government.” — Gavin Baker

“I do not want a few humans in control of intelligence.” — Gavin Baker

“This is an attempt by a specific political movement in San Francisco, moving in conjunction with one of the major political parties, to insert itself in the operations of every AI company, nominally in pursuit of a goal that is futile.” — Ben Thompson

“‘pacing the frontier’ in a world of geopolitical competition with China is stupid and not going to happen; that reality ought to end the whole argument.” — Ben Thompson

Charts and figures

From the issue’s “Posts of the Day.” Three other posts there are unrelated jokes and were not saved.

Minh Nhat Nguyen (@menhguin), Sep 13, 2026: the weekend’s turn summed up as a meme. The labs agree to pace, and the White House says no (the FT report in the timeline).

Corey Quinn (@QuinnyPig), Sep 12, 2026: the joke reply to the pacing announcement.

How it connects

• Dario Amodei: “We Must Pace the Frontier” is the precipitating event, and TBPN reconstructs the 48-hour timeline around it.
• AI Safety: the pacing debate is AI safety as governance (who slows down, by how much, and who verifies it), as opposed to alignment research.
• Tech Regulation and Regulatory Capture: Sacks’s “cartel” and antitrust-suspension objections, Baker’s predicted executive order, and the Lina Khan headline all frame pacing as a regulatory-capture question.
• Two Missing Pieces in the AI Safety Discussion: Noah Smith covers the same weekend from the pro-safety side. He accepts Thompson’s China objection and proposes a way around it (scaring the CCP with its own models) instead of treating it as the end of the argument.
• Niall Ferguson and Tyler Cowen — AIpocalypse Now: The Free Press’s same-morning take on the same consensus, pairing an alarmed essay (Ferguson) with a skeptical one (Cowen, “the AI slowdown is not what it seems”).
• The Dangerous Ideology Behind the AI Warnings: Shyam Sankar sharpens Thompson’s “specific political movement in San Francisco” into a named target, Effective Altruism.
• AI Is Already in Dangerous Hands: Ed Zitron’s same-week piece. Like Baker, he puts the danger in concentrated control rather than in rogue AI.
• Ben Thompson and Stratechery: Thompson’s “Pacing the Frontier, AI’s Digital Limits, AI Commissars” is the sharpest critique quoted. Compare his earlier Regretful Accelerationism.
• David Sacks: his “go ahead, you are the frontier” response, which Vinod Khosla quote-tweets in Smith’s piece too.
• AI Is Oil, Not God: Sacks’s duopoly framing matches Packy McCormick’s point that a frontier duopoly is what open-weights backers are fighting.
• Meditations on Moloch: Thompson’s view that no lab or country will slow while a rival keeps going is the coordination problem that essay describes.
• China: the hinge of the critique, and the subject of several of the issue’s headlines (China’s state paper, the WSJ).
• TBPN and Jackson Fordyce: publisher and co-author of the daily “Run of Show” newsletter.