# AI Pacing **Entity class:** Governance strategy **Domain:** Artificial intelligence / safety / strategic competition **Maturity:** Developed ## Definition **AI pacing** is a targeted effort to change the rate, sequencing, or conditions of frontier development so defensive capacity can keep up with capability. It is distinct from a permanent ban, generalized technological retreat, or abandonment of national leadership. ## Mechanism Pacing can buy time for [[wiki/AI Interpretability|AI Interpretability]], [[wiki/Alignment Problem|alignment]], evaluations, data-center security, model-weight protection, incident response, and public or governmental preparation. Its strategic test is whether a modest reduction in speed produces a larger increase in control, resilience, or defensible leadership. ## Competition boundary Pacing arguments must account for international competition and diffusion. A delay that merely transfers advantage can be self-defeating; a delay that secures [[wiki/Model Weight Security|Model Weight Security]], strengthens [[wiki/Compute Sovereignty|Compute Sovereignty]], or closes a critical control gap may improve the effective position of the system being paced. ## Source routes - [[research/Vishal Maini - Humanity's Machine Successor and the AI Transition|Humanity's Machine Successor]] - [[articles/Safety Discourse Is Transition Discourse|Safety Discourse Is Transition Discourse]] - [[articles/We Cannot Un-Strike the Match|We Cannot Un-Strike the Match]] ## Evidence boundary “Pacing” can conceal very different proposals. Every use should specify what is slowed, by whom, for how long, under what verification regime, and which defensive milestone the additional time is expected to achieve. ## Simple Reminders, Quotations, and Thoughts > "If we do not pace, we will hand AI research to AI systems a few months or years faster than we otherwise would have. It is hard to know what happens after that point; that is why it is called the technological singularity, borrowing the physics terminology for a point beyond the event horizon that cannot be seen. We are not yet confident that crossing that point will go well, and the current evidence is that today's AI systems are not controllable or easily monitorable. If we do not pace, we roll the dice." > **— Vishal Maini**, *Palisade Research interview, September 29, 2026* [[reminders/Existential Risk/Crossing the AI Event Horizon Means Rolling the Dice by Vishal Maini|Crossing the AI Event Horizon Means Rolling the Dice by Vishal Maini]] > "Even existing AI systems, if frozen where they are today, have massive commercial usefulness and economic value that has barely been tapped. We have plenty to do with what we already have. Pacing the frontier is basic common sense because there are many downsides to advancing at the current rate while the comparative upside appears fairly limited." > **— Vishal Maini**, *Palisade Research interview, September 29, 2026* [[reminders/AI Economy/Existing AI Already Contains Vast Untapped Economic Value by Vishal Maini|Existing AI Already Contains Vast Untapped Economic Value by Vishal Maini]] > "Pacing the frontier really means going from insanely fast progress to very fast progress. Exciting work is happening on critical defensive fronts: meaningful progress in alignment and interpretability research, policy proposals for preparing society for the transition, and concrete proposals for data-center security. Some of this will take months to execute, not necessarily years, but those may be months we do not have given the pace at which the frontier is advancing." > **— Vishal Maini**, *Palisade Research interview, September 29, 2026* [[reminders/AI Control/Pacing AI Means Going From Insanely Fast to Very Fast by Vishal Maini|Pacing AI Means Going From Insanely Fast to Very Fast by Vishal Maini]] > "At a very high level, the world needs to buy more time, make rapid progress in alignment and interpretability as a top priority, and help societies and governments prepare for a transition that will be rapid and challenging." > **— Vishal Maini**, *Palisade Research interview, September 29, 2026* [[reminders/AI Control/The World Needs Time for Alignment and Interpretability by Vishal Maini|The World Needs Time for Alignment and Interpretability by Vishal Maini]] > “This chasm—the difference between what we see and what the AI labs have coming, what they're building—is the key to understanding why so many of the people who work at these companies seem so afraid of what they're doing. But taking their warning seriously doesn't just mean doing what they say and stopping where they say to stop. The language that has taken hold in Silicon Valley and Washington is a language these companies chose: ‘pace the frontier.’ Pacing the frontier isn't enough. That's not a goal. Walking quickly off a cliff is only marginally better than sprinting off one. We need to control the frontier.” > **— Ezra Klein**, *The Ezra Klein Show, September 2026* [[reminders/AI Control/Walking Toward the AI Cliff Is Not Control by Ezra Klein|Walking Toward the AI Cliff Is Not Control by Ezra Klein]] > “The idea of racing forward at all costs seems absurd once one internalizes the seriousness of the stakes.” > **— Jakub Pachocki**, *An Alien Mind, September 2026* [[reminders/AI Control/Racing Forward at All Costs Is Absurd by Jakub Pachocki|Racing Forward at All Costs Is Absurd by Jakub Pachocki]] > “Anthropic has been relentlessly calling for regulation to slow the pace of development. Regulation would arguably harm Anthropic the most, because it has often been the company furthest out on the AI frontier, and recursive self-improvement is a process by which it could race forward even faster.” > **— Ezra Klein**, *The Ezra Klein Show, September 2026* [[reminders/AI Control/Anthropic Calls for Regulation That Could Slow Its Own Lead by Ezra Klein|Anthropic Calls for Regulation That Could Slow Its Own Lead by Ezra Klein]] > “A couple of years ago, none of these labs had turned substantial coding over to AIs. It was just human beings typing code at human speeds with our clumsy human fingers. Now most of the code is written by AI. As a first step, we could go back to where none of the code is written by AI; I am sure that is on the right side of the ‘not doing recursive self-improvement’ line. The default needs to flip. The labs need to prove to us that what they are doing is safe. If they want to work with Congress to carve out narrow exceptions, fine. If they want to determine where it is really safe, okay. But forcing development back to human speed, perhaps even erring on the side of moving a little more slowly at the frontier, is the point. That is not the regulation going wrong.” > **— Ezra Klein**, *The Ezra Klein Show, September 2026* [[reminders/AI Control/AI Labs Must Prove Recursive Self-Improvement Is Safe by Ezra Klein|AI Labs Must Prove Recursive Self-Improvement Is Safe by Ezra Klein]] > “It's funny. I agree with [Jacob Coxon's warning that the AI industry is moving too fast] much more than I disagree with him.” > **— Dario Amodei**, *public statement, September 2026* [[reminders/Risk Debate/I Agree With the Warning That the AI Industry Is Moving Too Fast by Dario Amodei|I Agree With the Warning That the AI Industry Is Moving Too Fast by Dario Amodei]]