Pausing AI ASAP is preferable to agreeing to pause at some future time

Connor Williams

Pausing AI ASAP is preferable to agreeing to pause at some future time — LessWrong

93 Pausing AI ASAP is preferable to agreeing to pause at some future time

8th Sep 2026

2 min read

93

It's much simpler, more straightforward, and more robust to pause as soon as possible than to 'agree to pause in the future, and then do it at just the right moment'. Several reasons:

Pausing doesn't happen in an instant. It will, in practice, likely take at least weeks if not months for an enforceable pause to be implemented once everyone agrees it is time. In that period, capabilities development will continue until the absolute last moment.
The shorter the runway from the present AI status quo to truly dangerous capabilities, the more fragile the pause will be. And the longer we wait, the more complex the necessary measures will be to enforce the pause and prevent rogue actors from covertly advancing the frontier, because we'll have less time to safely detect them and stop them.
Fewer points of failure. To agree to pause at some point in the future, you both have to get people to agree to that future goal, then when the time comes you have to get them to agree again. Each distinct time that everyone has to agree on something, there's an opportunity for failure/deception/defection.
Relatedly, if a bunch of bigwigs announce that we're agreeing to pause, but at some point in the future rather than right now, that gives lobbyists for the frontier AI companies, and for AI accelerationism in general, a useful signal to rev up efforts to lobby against a pause. It might prompt them to throw tens of millions, even hundreds of millions, into anti-pause lobbying and PR, whereas an immediate pause agreement may catch them off guard and deny them this opportunity.
This might likewise be a signal to frontier AI companies that they should borrow hard against future earnings, scale up staffing and compute as fast as possible, and do everything in their power to advance capabilities as much as possible before the pause. It might also motivate them to advance internal deployments even further ahead of external deployments than is presently the case.
Political capital is easier to build for the former than for the latter, especially beyond narrow technical circles, because it's a simpler, more intuitive, and more emotionally satisfying goal.
It's inherently difficult to know just 'when the right moment' is. There are error bars to this determination, and the longer we wait, the narrower the acceptable error bars become.

Frontpage

93

Pausing AI ASAP is preferable to agreeing to pause at some future time

New Comment

8 comments, sorted by

top scoring

Click to highlight new comments since: Today at 7:34 AM

[-]vals tutor1mo2814

Mostly meta comment, this is an example of a post where I'd benefit from a separate karma vs agree/disagree vote. It's a fine conversation opener and thus fine for positive karma, tho imo not as much as it currently has for how short-form quality it is, but I guess lots of people are upvoting because they agree and not because it's a high quality contribution that deserves 100+ karma and to be shown to everyone.

In practice I just strong downvoted as correctional for how upvoted it is because I think it's plausibly wrong and doesn't much support its case for someone who's not already convinced. It's kinda bad/worse for epistemics to list 10 qualitative reasons in one direction without listing the reasons in the other and doing a quantitative sizing them up. Policy Debates Should Not Appear One-Sided

[-]Ebenezer Dukakis1mo54

weeks if not months for an enforceable pause to be implemented

That sounds like a quite optimistic timeline for a major international treaty to be negotiated etc.

[-]Erez Abrams1mo3-1

I found this post initially confusing because it seems like you're contrasting "agree right now to pause right now" against "agree right now to pause at some unspecified future time", where from the title I expected the post to contrast "agree to pause right now" against "agree to pause at some particular future time". In particular, the first point in particular seems like an argument against "agree right now to pause right now" and in favor of "agree right now to pause at some particular future time". It would be helpful if you clarified your stance between all three scenarios (pause now, pause at specified future, pause at unspecified future) (either as a reply here or, ideally, by editing your post), because right now it's still unclear to me which of these arguments were intended to apply towards each scenario.

[-]momom21mo13

It seems to be clear that in practice, we cannot push towards pausing at a specific point in the future. There is no Schelling point nor a clear red line common for everyone.

[-]Erez Abrams1mo10

You say it's clear that in practice we can't do "specified future", but this is not at all clear to me. In particular, it seems clear that in practice we can't do "pause right now", since as far as I know there's currently no magic red button labeled "nondestructively pause specifically all training jobs" that the CEOs can ceremoniously press together at the drop of a hat; pausing training would require some internal time and effort, so something like "pause at the end of September" seems to me possible whereas "pause immediately" seems impossible. (If we then weaken "agree now to pause right now" to "agree now to pause as soon as possible", it seems like everyone has incentive to half-ass the pausing effort, and it would be hard to tell who is acting in good faith; the easy resolution is "agree now to pause within two weeks", which is "pause at specified future" again.)

With regards to pausing in the future: of course there is no clear Schelling point a priori, but the nice thing about Shelling points is that it's easy to manufacture sufficiently-common knowledge and therefore easy to generate a coordination point (which can then become Schelling by virtue of its salience); if Anthropic were to say "if OpenAI commits to pausing all frontier training runs by the end of September 2026, we also commit to pausing all training runs by the end of September 2026", then voilà, a Schelling point is born.

[-]_rpd1mo1-1

as a practical political matter, any global pause at a specified date within 6 months can be considered pausing "right now." pause on condition met, e.g., pause on evidence of RSI (however defined), might actually result in an earlier pause.

[-]neo1mo20

This might likewise be a signal to frontier AI companies that they should borrow hard against future earnings, scale up staffing and compute as fast as possible, and do everything in their power to advance capabilities as much as possible before the pause.

I wonder if frontier labs would be able to continue rushing forward in the face of mounting public pressure against the race to ASI. If modeled as incredibly cohesive entities controlled by a few key decision-makers, maybe. But it seems like there is enough internal division about safety to make this somewhat difficult? I guess I struggle to imagine capabilities workers being capable of such anti-social behavior.^[1]

^{^}
"Anti-social" here meaning willing to accept a loss of social status to keep advancing capabilities. Which I expect to be rare, since I model most humans as primarily status-seeking.

[-]cguth71mo10

Also there will still be improvements even if we never have model better than current frontier. Harnesses, system prompts, apis that work well with agents, company integrations with AI and people's skill at using them will all continue improving. It's really hard to know how much additional capabilities could come from things listed or new similar innovations on top of them. Probably not relevant for the RSI takeover concerns but for the more banal concerns like bioterrorism and hacking and severe economic disruption.

Moderation Log