We should act as if we live at the hinge of history

By OscarD🔸 @ 2025-10-18T22:06 (+11)

The hinge of history hypothesis (3H) suggests, roughly, that there will be no future time at which altruistic and philanthropic action can have a higher impact than now. In section 1, I review some of the existing literature (so familiar readers could skip this) and in section 2, I propose a novel (to my knowledge^[1]) pragmatic argument for why we should act as if 3H is true. In short: false negatives are costlier than false positives in this case, so we should act based on 3H even if our credence in it being true is <50%.

1. Review of the 3H debate

There are some intuitively compelling reasons to suggest 3H is true (MacAskill 2022):

• Time of perils. The chance of an existential catastrophe this century is, putatively, disturbingly high. But this will not remain the case: either we will perish in short order, or reach a state of greater security and survive into the far future. Therefore, insofar as reducing X-risk is one of the most important longtermist goals, that goal can be pursued uniquely well now, making this the hinge of history.
• Pace of change. More broadly, the pace of technological change and economic growth over the last century or two has been incredibly rapid by the standards of human history. But such rapid growth cannot continue for more than a few millennia, if that. And ‘hingeyness’ is likely associated with times of rapid change, as there are more choice points and path-dependencies in a fast-moving civilisation than a static one.

Contra this intuitive case, MacAskill (2022) mounts two opposing arguments:

• Base rates. If the future contains very many people, the proportion of people who ever live who are alive today, or anytime this century or so, will be minuscule. Therefore, the base rate of any given person living at the hinge of history is very small. On a plausible view of anthropics, this means our prior on 3H should be very small. And plausibly, we don’t have sufficiently strong evidence to move so far from this prior to think that 3H is likely.
• Induction. It seems (in retrospect) that people’s influence from a longtermist perspective has been fairly monotonically increasing over time, as wealth, knowledge, and technology grew. For instance, hunter-gatherer societies had very little ability to plan for or influence even the fairly near-term future, and agrarian societies had only slightly more ability to do so. Scientific-industrial societies have far more ability to make predictions about the future, understand the natural and human world, and make targeted interventions. And to this day, we seem to have been coming up with new crucial considerations that increase our chances of influencing the far future positively. So by induction, it would be surprising if we live at just the moment when the graph of hingeyness over human history is just at the point of peaking, and will fall precipitously in future generations.

Mogensen (2024) adds to this by challenging MacAskill’s specific formalisation of 3H. In particular, MacAskill says we live at the hinge of history if we are “among the very most influential people ever, out of a truly astronomical number of people who will ever live.” But then, Mogensen notes:

On MacAskill’s interpretation, if we knew we lived at the ‘hinge of history’, zero urgency would attach to lowering the risk of extinction this century, since 3H entails that humanity will not become extinct for a very long time.

To avoid this incongruity, we could instead specify in the definition that there are an astronomical number of people in expectation. But this then removes much of the force of MacAskill’s base rates argument, since, if we are indeed in the time of perils, there is a decent chance we are among the last generations to live, and therefore at the hinge of history because we were (putatively) the last generation with a chance of preventing extinction. I am uncertain whether MacAskill’s base rates argument can be rescued, but it at least seems clear that it requires additional effort to reformulate more carefully.

Mogensen also worries that the inductive argument fails to account for the fact that one way to be highly influential is to invest resources for the future. If investment to donate later counts as influentialness, then plausibly earlier times were actually more influential, provided people practised patient philanthropy. I agree with Mogensen that the argument requires some finessing about how to classify investment, but I think some version of the argument likely carries through.^[2] That said, it seems to me that this inductive argument is far less ambitious than the base rates argument, as at most there are a few datapoints of past eras where influentialness has increased monotonically. There is little reason to think the year 700CE was more influential than 600 CE, so we are just left with information age > industrial age > agricultural age > prehistory in terms of hingeyness.

More recently, Häggström (2025) revisits the base rates argument, concluding that even a very low prior probability is not decisive, given the strength of our evidence for 3H. In particular, some events that have occurred or are likely to occur soon – such as humanity first venturing into space, or first developing advanced AIs – may be quite likely to occur near the hinge of history given how momentous they are in the story of civilisation. And the prior probability that we are alive so nearby in time to these momentous changes provides very strong evidence for 3H, enough to undo most of the tiny prior, perhaps. Indeed, a similar point was made earlier by Shlegeris (2020), who notes that the main thing we should be surprised about (on anthropic grounds) is how early we are in history, if the future is large, but conditional on being early, being at the hinge of history isn’t very unlikely.

Separately, Häggström also notes, reasonably, that anthropic reasoning is difficult and we should not be at all confident we have even the basics settled for how to do so properly. And this provides some upper bound on how compelling the base rates argument should be.

2. A pragmatic argument for acting on 3H

Here is an argument for acting as if 3H is true (without necessarily believing 3H to be true):

1. If 3H is true but we invest patiently, this false negative is catastrophic, as it could significantly increase the chance of an existential catastrophe or missing out on a flourishing future.
2. If 3H is false but we act urgently, this false positive is far less bad, as we will have many years (maybe millions) later in which to invest resources for the real hinge of history.

      C. Therefore, it is prudent to act as if 3H is true, even for credences <50%.

There is a fundamental asymmetry here, where if 3H is true, our actions matter enormously (by definition), whereas if 3H is false, our actions matter rather less. And since 3H recommends acting urgently, the stakes are higher in that case, so we should act urgently. Note, there is some subtlety about how ‘investment’, whether financial or otherwise, should count here, which I won’t go into.^[3]

Premise 1 seems obvious: if we let the hinge of history pass us by without acting on it, we will (by hypothesis) never have as great an opportunity to create longtermist value. And our future opportunities may not just be a bit worse from an impact perspective, but orders of magnitude worse (if we are trying to spend philanthropic resources after we are already well on the road to a near-best future) or nonexistent (if humanity goes extinct before we spend down our resources).

Premise 2 is less clear-cut. For instance, if the hinge of history is in a few decades, but by acting on 3H we spend down our wealth, political capital, and other resources to a low level before then, this could be very bad. Conversely, if the hinge of history is in a thousand, or especially a million, years, our actions now probably have fairly little impact on how well humanity performs in the real hinge of history. This is both because it is hard to successfully steward resources over longer timescales, and because even if we spend down resources for a decade or a century, we have longer to realise our error, and become patient philanthropists with plenty of time to go. This does seem to limit the conclusion somewhat – clearly we should not spend all our resources within one year, even if we are urgent longtermists. But I think the general point stands that, in expectation, a false positive on 3H is not so bad.

This pragmatic argument is reminiscent of the Presumptuous Philosopher, who is supremely confident that we live in a large world, because the fact that we exist is evidence that there are many observers. Or we could imagine an alternate version where, aside from anthropic reasoning, the philosopher now acts on the belief that we live in a large world, because then our actions have vastly more causal and evidential impact. In the 3H case, this line of argument suggests we don’t need to worry as much about carefully thinking through civilisational path dependencies and hingeyness, and can just jump straight to acting on 3H because the stakes are higher. This feels a bit like cheating.

I don’t have a very satisfying answer; I think we can just bite this bullet. Our actions need not be based on what is most likely but rather on the combination of what is likely and what is most high-stakes (consider buying insurance).

Another weakness of my argument is that it is not (yet) quantitative. Ideally, we would know how small our credence in 3H should be for us to stop acting on it. Clearly, 10% is very different from 1% or 0.1%. I don’t have the analytical tools to address this well, but I would be excited to see someone try. I think the task would involve:

• Assigning credences to different possible trajectories of how hingeyness will change over time (based on e.g. different possibilities re the time of perils).
• Making some assumptions about growth rates, rates of return, rates of value drift or other resource loss, etc.
• Running Monte Carlo simulations of a large number of possible futures, for each of various actions we could take today (perhaps quantified by the rate at which we spend down philanthropic resources) and then assessing the best action on different ethical and risk-aversion views.

Probably the best place to start would be taking Phil Trammel’s existing patient philanthropy model and coupling it with some hinge of history modelling. But for now, I will settle for raising this pragmatic argument for acting on 3H qualitatively, leaving a more rigorous quantitative version to future work.

1. ^{^}

    Note: this seems like quite an obvious point, so quite possibly it is not original to me, but I don’t remember seeing it written up anywhere.

2. ^{^}

    For instance, if we would have preferred our past selves (or our forebears) 1, 5, 10, 50, 100, etc years ago to invest money for us to donate now, than for them to have donated it directly, then it would indeed be somewhat surprising if now was the first time that donating directly was better than investing to give later.

3. ^{^}

    If investing for later counts towards the current time’s hingeyness, then now may be the hinge of history on a wider range of views about how the future will go. But 3H being right then tells us rather little about what we ought to do.