What this covers

William MacAskill, an Oxford philosopher and founder of the effective altruism movement, discusses longtermism with host Russ Roberts in conversation drawn from MacAskill's book What We Owe the Future. The core argument holds that future people deserve equal moral weight in our decisions today because they could vastly outnumber us, their interests are genuinely at stake, and our choices meaningfully shape their circumstances. MacAskill distinguishes his view from an implausibly demanding moral absolutism: he argues that even modest present sacrifices are easily justified on multiple ethical frameworks when the scale of future populations is considered. The conversation ranges across philosophy, economics, history, and institutional design, grounding abstract moral claims in concrete cases and tradeoffs.

The discussion treats Contingency of Moral Progress as central—the claim that abolition, democracy, and other moral advances were not inevitable historical forces but depended on specific people and cultural shifts. MacAskill illustrates this through Benjamin Lay and other Quaker moral entrepreneurs whose heterodox conviction and dramatic persuasion helped end slavery in Britain and globally. This grounds a broader point: while technological innovation appears nearly inevitable across cultures, moral change remains fragile and contingent, making moral agency now consequential. The conversation also examines harder terrain: whether future people will reliably be richer (MacAskill argues the claim applies only to marginal harms, not catastrophes), how Thinking at the Margin can obscure tail risks, tensions between global coordination and value lock-in, and whether having children can be morally defensible. MacAskill defends Moral Impartiality—that location in time, like location in space, does not reduce moral concern—while acknowledging ongoing philosophical puzzles about why anyone should care beyond themselves.

Sharpest takeaway

MacAskill argues for longtermism—that positively influencing the long-term future is a key moral priority because future people count morally, could be vastly numerous, and we can meaningfully affect their lives—while distinguishing this from extreme demandingness and emphasizing that the path of moral and civilizational progress is contingent, not inevitable.

  • Future people have equal moral status, and given their potential numbers, even modest present sacrifices are easily justified.
  • The argument that future people will be richer applies only to marginal harms, not to catastrophes, enslavement, or non-existence.
  • Moral progress (e.g., abolition of slavery) was contingent on cultural change and moral entrepreneurs, not inevitable economic or historical forces.

The argument · threads7 threads · 31 claims
0.79

Societies must protect moral diversity and reasoning over centralization to preserve space for moral progress.

4 pointscentrality 5/5
  • Rather than imposing a narrow totalizing moral view, we should build a society with great diversity of moral views and institutions in which views can debate, reason, and experiment so that the best moral ideas win out on their merits rather than via conquest, since we are probably still far from the morally best view.

    “what we want to do is build a society that can have a great diversity of moral views and a kind of culture and institutional setup such that those views can debate and reason and experiment. And, we can learn over time which the right moral view is. And so, the kind of best ideas went out on their merits rather than via conquest.”

  • While many catastrophic risks (nuclear war, dangerous technology, carbon emissions) push toward greater global centralization to solve coordination failures, excessive centralization carries its own catastrophic risk of stalling moral progress—through value lock-in under a world dictatorship or gradual homogenization where people stop trying to improve morally—and good institutions must thread this needle.

    “there are risks on that end, too, where greater centralization could mean we stall moral progress. In the worst case, the simplest case, you've got a world government; it's a dictatorship; there's another ideology that's locked in forever.”

  • A society aiming at moral progress is undermined when its most influential people—politicians—are expected to lie, willfully neglect the truth, or make technically true but misleading statements; this should be treated as an absolute scandal so that powers of non-rational persuasion are muted and argument and reason prevail.

    “we kind of expect politicians to lie. If they do, maybe they get a little bit of criticism... I'm like, this is terrible. We're trying to have a world where we get to better views, and having the most influential people in society... outright lie, or willfully neglect the truth.”

  • The right epistemic stance for a moral agent is 'strong opinions, weakly held'—holding and defending views firmly while remaining willing to switch immediately upon a decisive counter-argument, and pursuing change through rational persuasion and incremental steps rather than violent revolution.

    “This motto that I'm sympathetic to, which is strong opinions, weakly held... if I have an argument that I think is decisive against my view, I switch... I want to achieve that end via rationally persuading people, and perhaps that involves going by a number of incremental steps.”

0.79

Future people matter equally to present people; catastrophic harms to them are morally urgent regardless of timing.

4 pointscentrality 5/5
  • If you know with certainty that a harm will occur, it does not matter morally whether it happens tomorrow, in a decade, or in a thousand years—location in time, like location in space, is morally irrelevant to the badness of harm, because morality requires impartially taking seriously the interests of anyone you affect.

    “supposing we know that that will happen, does it really matter whether that person cuts themself tomorrow or in a year's time or a decade? or even if it was 100 years or 1,000 years? And, I think intuitively: no. As long as you're certain that that's going to happen, then harm is harm.”

  • People in the future have the same moral worth or moral status as people in the present, though we may have additional special moral duties toward those we have relationships with; the impartial baseline is that everyone counts equally.

    “I think people in the future have the same moral worth or moral status as people in the present. I do think there are additional reasons that are not about people being present, but are about relationships we have with people in the present.”

  • Because all people count equally, saving ten lives is morally more important than saving one; the interests of larger numbers of people carry proportionally greater moral weight, which is why the sheer potential number of future people raises the stakes of longtermism.

    “if I can save one life or save ten lives, it's more morally important to save the ten. And that's, again, because all people count equally, so the interests of ten are just more important than the interests of one.”

  • The argument that future people will be richer only applies to marginal financial harms; it does not apply to catastrophic harms like enslavement, death, or never coming into existence, where prior wealth is irrelevant—so longtermist concern focuses on catastrophes and value changes, not marginal differences in wealth.

    “this argument, that future people will be richer than those today, that only applies to what economists call marginal harms... Is that person in the future enslaved? It really doesn't matter where--how rich they were beforehand... If they're now enslaved, that's not like a marginal harm.”

0.65

Moral change is contingent and driven by moral entrepreneurs using reason and persuasion, not inevitable economic or technological forces.

4 pointscentrality 4/5
  • Technological innovation is largely inevitable because intense incentives across many cultures and moral views drive humanity to invent things eventually, but moral change is much less inevitable—in many cases it could go either way—which is why moral entrepreneurs who shift the trajectory matter enormously.

    “the intense incentives across many different kind of cultures and moral views you could have for technological innovation means that I think by and large we'll get there eventually. With a case of moral changes, though, I think that's much less obvious; and at least in many cases, things could go either way.”

  • The abolition of slavery was largely a matter of contingent cultural change rather than inevitable moral progress or economic forces; Quaker-Enlightenment thought convinced the British elite and public to end slavery and pressure other powers, taking forced labor from near-universal in 1700 to about 0.5% of the population and illegal everywhere today.

    “prior to learning about this, I would've thought, 'Yeah, this is inevitable. It's either just like the inevitable march of moral progress, or it's just the result of economic changes.' And, I no longer think that's the case. I think it was largely a matter of cultural changes.”

  • Benjamin Lay, an 18th-century Quaker dwarf, was a paradigmatic moral entrepreneur—a heterodox 'moral weirdo' who used dramatic guerrilla theater and major personal sacrifice to campaign against slavery, contributing to a Quaker and Enlightenment movement that helped end slavery, illustrating that we should aspire to such conviction-driven moral agitation.

    “Benjamin Lay is a Quaker. He's a dwarf. He refers himself as 'Little Benjamin who beat Goliath'... In his most famous stunt he brought a Bible that was filled with fake blood to the 1738 meeting of the Quakers, and said it was as great a sin to keep enslaved people as to stab the Bible.”

  • A key difference between a model moral entrepreneur and a dangerous fanatic is means: Benjamin Lay used reason, empathy, and peaceful protest, whereas conquest and violence are illegitimate because they are not methods of arriving at correct moral views—though MacAskill concedes the line between rational persuasion and brainwashing is genuinely hard to draw.

    “an enormous difference between Hitler and Benjamin Lay is a difference of means. So, Benjamin Lay was agitating. He was making arguments, he was engaging in peaceful public protest. It was via the power of reason and empathy.”

0.65

History was overwhelmingly dark; recent moral progress like slavery abolition shows cultural change is possible but fragile.

4 pointscentrality 4/5
  • Human history was overwhelmingly dark: in 1700 the majority of people were in some form of forced labor, the world was extremely patriarchal, and most lives involved enormous suffering and ill health—whereas today, even on a broad definition, forced labor is about 0.5% of the population and illegal everywhere.

    “if you just look at what is human history like--and it's a dark place. People in the late past had miserable lives... the majority of people were in some form of forced labor is my best guess.”

  • The most important long-term aspect of civilizational collapse may be the values a recovered world holds: the prevailing egalitarian, liberal, democratic culture is at least somewhat lucky and not guaranteed to re-emerge, so a recovered world at today's technological level could plausibly feature widespread slavery or authoritarianism, making it considerably worse indefinitely.

    “if you tell me that, kind of, there's a catastrophe, the world comes back, we get to this level of technological development, but slave-owning is very widely widespread, or the large majority of countries in the world are authoritarian rather then democratic, I'm not the least surprised.”

  • The reassurance that 'we've been doing well and things keep improving' is brittle because it ignores near-misses: the probability of an all-out US-USSR nuclear exchange may have been around one in three, and in the world where it happened we would not be congratulating ourselves on progress, so we should drive such catastrophic risks as low as possible.

    “What was the probability that the United States and USSR would have an all-out nuclear exchange? And, if I had to guess, it'd be something like one in three... I would like us to be in a world where that risk of all-out nuclear war was not one in three: instead it was more like 1%, or 0.1%.”

  • The 20th century gave warning signs of catastrophe on both the technological side (nuclear weapons) and the value side (totalitarian regimes arising even out of democracies); a different roll of history could plausibly have seen Nazi fascist values establish a long-lasting regime.

    “we've seen the first warning signs in the 20th century on both the value side and the technological side. Technological side, we saw nuclear weapons. On the value side, we saw totalitarian regimes, even a rising out of democracies.”

0.65

We spend far too little on long-term future guidance relative to catastrophic risks we create; modest resource sacrifice is easily justified.

4 pointscentrality 4/5
  • Because so little of world GDP (perhaps 0.1% or 0.01%) is currently spent on positively guiding the long-term future, we are far from any extreme demandingness margin; sacrificing even 1% of resources for future benefit is easily justified on a wide variety of moral views, not just utilitarianism.

    “in terms of effort that we spend really trying to think about the long-term impacts of our actions... how much of world GDP is that? I don't know: 0.1%, 0.01% or something? It's very low. And so, on the current margins, if we get to 1%, I would be over the moon.”

  • Helping people today already benefits future people because innovation, better institutions, and improved moral culture carry forward, though MacAskill notes it would be surprising if such ordinary helping were the best way to help the future given that some present activities (like CO2 emissions and risky biotechnology) actively harm it.

    “when we innovate or build better institutions or have a better, more moral culture, all of those things do benefit future people, too. It would be surprising though if that was the best way of helping people.”

  • Contrary to the unargued assumption in economics that future people will be much richer, there is meaningful uncertainty—MacAskill estimates roughly an 80% chance the year 2300 is better off but a 20% chance it is worse, due to risks of catastrophe, war, engineered pandemics, or civilizational stagnation and decay.

    “To take the year 2300, I would say it's like 80% that are better enough than us, and maybe a lot better off... However, 20% chance that actually perhaps they're worse than us.”

  • Just as we have some concern for animals (we treat pets well) yet still keep roughly 80 billion farm animals in horrific conditions because our concern is far less than warranted, we have some concern for future people but inflict unnecessary harm because that concern falls far short of what is morally appropriate.

    “pets get treated pretty well: The 80 billion animals that are kept almost all in horrific conditions and then slaughtered, they have really terrible lives. And, I think there's an analogy between that and future people.”

0.52

Having children and moral progress depend on conscious experience, not preference satisfaction or narrow economic measures.

4 pointscentrality 3/5
  • Having children can be a morally good act, not a bad one: the carbon harm can be offset cheaply—even a thousand times over—by donating roughly $1,000 to effective climate nonprofits (about $1 averts a ton of CO2), and the carbon ledger ignores the immense positive contributions of a flourishing person to society, innovation, and moral progress.

    “you can nullify that harm by offsetting. And, in fact, you can nullify it 100 times over... By donating to extremely effective climate nonprofits, you can avert an expectation of a ton of CO2 for about a dollar.”

  • Economists assume a preference-satisfaction theory of wellbeing (what's good is getting what you want), but MacAskill prefers a theory where wellbeing is a matter of positive versus negative conscious experiences; on the latter view, people's preferences (e.g., against dying) only imperfectly track their wellbeing because evolution would bias us toward survival regardless of experiential quality.

    “economists normally assume what's known in the philosophy literature as a preference-satisfaction view, where what's good for you is getting what you want... My preferred view of wellbeing is that wellbeing is just a matter of conscious experiences.”

  • Today's high material standard of living and moral progress are in significant part a numbers game—the result of having had so many people contributing net-positively; had only half as many people ever lived, the average person would likely be a farmer working 12 hours a day in some form of forced labor with little freedom.

    “the fact that we have a world today where we have a high material standard of living and that we have made some moral progress, that's in significant part a numbers game: the fact that we've had so many people who have contributed in a net positive way to society.”

  • In countries like the United States, people end up having fewer children than they say they would like (around 1.8 versus a stated desire of about 2.6), a gap Roberts notes is between stated and revealed preference rather than necessarily true unmet desire.

    “at the moment in countries the United States, people have less kids than they want to have. I think they want to have, like, 2.6 on average, and they have 1.8”

0.44

Morality has no ultimate non-circular justification; impartial concern for others rests on bedrock reasons like suffering's badness.

5 pointscentrality 2/5
  • The question 'why should I be moral?' has no ultimate non-circular answer, but neither does 'why should I care about myself?'—both eventually bottom out in bedrock reasons (suffering is bad, happiness is good); the egoist who rejects reasons toward others has no privileged ground, since the same 'why should I care?' regress applies to self-interest.

    “this is one of the deepest questions in philosophy, which is: Why ought I to be moral? And, I think ultimately there's no non-circular answer... there's an equal argument for: Why should I care about myself?”

  • Appealing to God does not escape the 'why should I care?' regress, because one can always ask why one should care about what God wants or fear hell—the latter merely reduces to the self-interest question—so the theist simply adds one more explanatory level to which the same skeptical question applies.

    “Why should I pick up the glass? And, you say, 'Well, ultimately, because God wants you to.' And, I say, 'Well, why should I care about what God wants me to do?'... at some point you are just throwing a line or the theist is just giving one additional kind of level of explanation. But, the why question can be applied to that, too.”

  • Religion functions like a social technology, and the key innovation was 'Big Gods'—moralizing gods who watch you even when alone and care what you do—which reduces free-riding and cheating when widely believed; the post-religious era (the world is only ~16% atheist/agnostic) is so young that resurgent free-riding is a genuine worry.

    “the thing that appeared was many different religious traditions, what's called Big Gods. These are Gods who are watching you while you're alone... If everyone believes that, then you get a lot less free riding. You get a lot less cheating.”

  • Strong cultural and evolutionary prohibitions against taking one's own life are weak evidence that lives in the past were genuinely bad: if a subspecies ended their lives at high rates it would be selected against, and major moral traditions may have evolved strong suicide prohibitions precisely because, absent them, rates would have been much higher given how bad lives were.

    “it's pretty notable that in kind of major moral traditions, there's these very strong prohibitions against taking one's own lives. Why is that?... maybe people's lives in the past were just really bad, and rates to take ones own life would've been very high... were it not for those prohibitions.”

  • Humanity would probably recover even from a catastrophe killing 99% of the population, because historical catastrophes (Black Death, Hiroshima/Nagasaki) show remarkable resilience, vast preserved knowledge in surviving libraries would persist, existing tool prototypes would speed reinvention, and no clearly bottlenecking exhausted resource has been identified.

    “if you look at enormous but still smaller-scale catastrophes, like the Black Death in Europe, or even the bombings of Hiroshima and Nagasaki... you see people having remarkable resilience... a second reason... is just how much knowledge would be preserved... tens of thousands of libraries in locations that wouldn't be threatened.”