What this covers

Peter Singer and Tyler Cowen debate the reach and limits of utilitarianism in this philosophical conversation. Singer defends a maximalist position: moral reasoning should aim at universal happiness across all sentient beings, without special loyalty to one's own species. Cowen, who describes himself as "two-thirds utilitarian," presses back on whether the theory can actually do the work Singer claims. Their exchange moves across practical ethics, the nature of moral reasoning under uncertainty, and the institutions that shape how moral ideas spread in the world.

The conversation covers factory farming and animal ethics, where Singer argues the case for utilitarianism appears strongest, and broader questions about how to live a meaningful life. They examine the paradoxes that arise from strict utilitarianism—most notably Singer's claim that he would side with happier aliens if a common metric of happiness existed—and Marginalism's role in moral practice. Singer concedes that while utilitarianism holds at the theoretical level, practical obligation often requires following Rules of Thumb under Uncertainty, citing Sam Bankman Fried's overconfidence as a cautionary case. They also debate Effective Altruism's tilt toward long-term existential questions, Singer's evolving views on Meat-Eater Problem and progress in animal welfare, the role of tenure in protecting controversial scholarship, and why contemporary academic philosophy often feels narrow. The discussion shows where sentience serves as a threshold for moral concern and where calculation becomes intractable—such as evaluating whether humanity should ever have settled the New World.

Sharpest takeaway

Singer defends a thoroughgoing utilitarianism—loyalty to universal happiness over species partiality—while Cowen presses that the theory works only at the margin and that justice, partiality, and an underdetermined science of sentience expose its limits.

  • Singer rejects species loyalty and would side with happier aliens if a common metric of happiness exists
  • Both agree large total-comparison questions (e.g. settling the New World) are intractable
  • Singer concedes utilitarianism's practical force is strongest in clear cases like factory farming

The argument · threads8 threads · 34 claims
0.85

Sentience and capacity for suffering, not species membership, determine moral status and deserve equal consideration.

6 pointscentrality 5/5
  • A rich and fulfilling human life can reasonably be said to contain more happiness than the life of a cow or sheep; animals are not 'equal' in the sense of their lives containing equal value with humans, even though their interests deserve consideration.

    “I think it’s reasonable to argue that a rich and fulfilling human life contains more happiness than the life of, I don’t know, a cow or a sheep.”

  • Singer's willingness to side with happier aliens rests on the existence of a common metric of happiness; if there were no common metric, he would not know what to do—but this is not like cheering for a football team, because everything is at stake rather than mere tribal loyalty.

    “If there’s no common metric, I don’t really have an answer, or I would try to use the metric of overall happiness.”

  • Singer claims no particular loyalty to members of his own species above loyalty to the general good; if wiser, impartial aliens correctly calculated that a better world for everyone requires humans to go, he would side with them, because moral concern tracks universal happiness and well-being, not species membership.

    “I don’t claim to have any particular loyalty for my species rather than the general good.”

  • If intelligent sentient aliens exist elsewhere but can never reach us, the extinction of humanity remains the same loss—a loss of a certain amount of happiness—because the loss is absolute; only if surviving sentient life could regenerate and repopulate would extinction be worse than mere loss of the species, and there is nothing special about the zero point of a universe empty of sentient life.

    “the extinction of our species is still just the same loss. It’s a loss of a certain amount of happiness, and the fact that it’s less of a proportion of the happiness... doesn’t mean that it isn’t the same loss.”

  • Current large language models are not sentient and so do not count morally; sentience is essential for a being to count, and although the science of consciousness is incomplete and badly understood, Singer is not convinced that this incompleteness would lead us astray in making moral decisions based on sentience.

    “There’s something about sentience that is essential for a being to count, but we actually know very little about consciousness.”

  • Despite low fertility rates in secular wealthy nations, Singer supports secularism because it seems to him true—there is no divine being—and he cannot adopt belief in a divine being to fool people into having children; instead he hopes for ways to produce enough children for good long-term outcomes, such as baby bonuses, while accepting the risk that secularism could shrink if religious people out-reproduce.

    “one reason for supporting secularism is that it seems to me to be true. I don’t think that there is a divine being.”

0.80

Factory farming inflicts indefensible suffering vastly outweighing any human benefit and represents utilitarianism's clearest application.

5 pointscentrality 5/5
  • The suffering inflicted on animals in factory farms is indefensible because it causes far more suffering than any good it does for humans—arguably a net negative for humans too—making it one of the clearest cases where utilitarian calculation can be confident.

    “clearly, the suffering we inflict on animals in factory farms is indefensible. We cause far more suffering than any good we do for humans.”

  • Progress on animal ethics has been slow primarily because humans are conservative about what they eat—few are willing to think ethically and openly about their food for fear of having to change, eat differently, or implicitly criticize friends and family—and without changing what they eat people won't fundamentally change what they think about animals.

    “Relatively few people are prepared to really think ethically and openly about what they eat. They’re worried about having to change that, eat something different, maybe stand out from their friends and family by not eating meat”

  • To help other sentient beings one can: stop eating nonhuman animals (especially animal products from factory farms); support organizations combating factory farming, identified through Animal Charity Evaluators; and get political by telling representatives at all levels that animals matter and that one will vote for candidates with strong animal-welfare policies.

    “Firstly, we can stop eating nonhuman animals... Secondly, we can support the organizations that are trying to combat factory farming... Thirdly, you can get political about this.”

  • Being fully Darwinian is compatible with holding that humans evolved a capacity to reason, and that reason can lead to conclusions that influence behavior and extend a more universal ethic to a wider group of beings than kin, family, and cooperative partners—at least for some people some of the time.

    “we are a being who’s evolved the capacity to reason, and reason can lead us to conclusions which can influence our behavior. I don’t see that as being anti-Darwinian”

  • To live a happier life one should aim at making one's life fulfilling and meaningful—contributing to making the world better rather than just earning more money and consuming—because psychology research shows people who are generous and whose lives are in harmony with their values find their lives more rewarding and satisfying.

    “there’s a lot of psychology research showing that people who are generous and are thinking of values... do actually find their lives more rewarding and more satisfying.”

0.66

When confidence in utility calculation is low, adherence to principles and rules of thumb produces better outcomes than direct utility pursuit.

5 pointscentrality 4/5
  • The paradox of accepting a 51/49 double-or-nothing gamble to double Earth's population reveals a problem with maximizing expected utility rather than a flaw in utilitarianism per se; in the real world you cannot keep doubling utility infinitely, so the expected utility of repeated gambling is actually low and you shouldn't play double or nothing.

    “I think it says more about gambling than it does about being utilitarian.”

  • Although the fundamental obligation is to maximize utility, when we cannot be confident in our calculations we are more likely to make mistakes by aiming directly at utility, so our practical obligation may be to conform to principles or rules of thumb; this depends on confidence in one's own judgment, and Sam Bankman-Fried's mistake was being too confident he could fix things while violating basic rules like not stealing clients' money.

    “we should follow rules of thumb sometimes, when we can’t be sure of what’s the right outcome”

  • Even if one would ideally program a driverless car as a pure Benthamite, given that such programming would be shut down by law, public opinion, or the manufacturer, the right choice is to program it to produce the best consequences that would not be prohibited—making compromises to produce the most good achievable in the actual circumstances.

    “I would program it to produce the best consequences that would not be prohibited by the government or the manufacturer.”

  • Legal abortion does not straightforwardly reduce utility even if it lowers average births, because many abortions terminate poorly timed pregnancies while couples still have their planned number of children, and the children born to mothers who didn't want them at that time may be less happy and the parents more stressed; additional considerations like a child's lifetime greenhouse-gas consumption and habitat deprivation further complicate the claim that another child is a net positive.

    “those babies would be born to mothers who did not want to have them at that particular time... the child might be a less happy child, and the mother or the parents might also be much more stressed”

  • Singapore is a strong candidate for the most utilitarian government toward its own citizens, though Australian governments are also reasonably utilitarian; differences in measured happiness relative to income may owe to factors utilitarianism cannot control, such as Australia's better climate and natural assets versus Singapore's hot, humid environment.

    “I think the Australian governments are actually reasonably utilitarian... but Singapore’s a strong candidate.”

0.61

Maximizing utility remains the fundamental moral obligation, reducible to consequences even where non-utilitarian intuitions arise.

4 pointscentrality 4/5
  • At the theoretical philosophical level Singer maintains that retributive and other moral intuitions are all explicable in terms of utility, even though he admits to having non-utilitarian intuitions—declining Cowen's invitation to become a 'two-thirds utilitarian' who lets in an independent, non-reducible reason of justice.

    “At a theoretical philosophical level, I’m going to say no, it’s all explicable in terms of the utility. But as I said, that doesn’t mean that I don’t have nonutilitarian intuitions about it.”

  • As a now hedonistic (rather than preference) utilitarian, Singer holds that the only reason to pay attention to the wishes of the dead is that the living will then have more confidence their own wishes will be respected; preference utilitarianism faced the problem that, if the past accumulates and population shrinks, the living would be overburdened satisfying past preferences.

    “The only reason to pay attention to the wishes of the dead is that the living will have more confidence that their wishes will be respected.”

  • For a utilitarian, saying justice was done to Nazi leaders means punishment serves an expressivist function—condemning a terrible act so people learn and are educated from it—and a deterrent function for others who might otherwise expect to get away with such evil; this is compatible with utilitarianism without requiring retributive justification.

    “we want to punish people who do great evil as a way of indicating... an expressivist function, that this was a terrible thing that they did... And also a deterrent function”

  • Parfit's late attempt to reconcile consequentialism, Kantianism, and contractualism into a 'triple theory' is partly a failure because he became so committed to showing morality is objective—since otherwise nihilism would be true and his life a waste—that this commitment led him to accept arguments he would otherwise have rejected; moreover he reconciled rule consequentialism, not act consequentialism, and Singer dislikes the resulting form of consequentialism.

    “he became so committed to the idea that he had to show that morality is objective — that otherwise nihilism is true, nothing matters, and his whole life would be a waste — that that influenced his acceptance of arguments that otherwise he would not have accepted.”

0.51

Effective Altruism should maintain focus on reducing present suffering rather than shift excessively toward speculative long-term existential scenarios.

2 pointscentrality 4/5
  • There is too much emphasis on existential risk from AGI, in part because many people drawn to AI enjoy the intellectually appealing problem of aligning superintelligent AGI with human values—a more interesting problem than reducing factory-farm suffering or extreme poverty—which biases attention toward it beyond what is justified.

    “a lot of the people in AI are people who like these kinds of problems. How are you going to align super-intelligent AGI with human values? That’s a really interesting problem”

  • Effective Altruism has moved too far toward very long-term thinking (the next million or billion years); while preventing extinction matters, there is uncertainty about whether we can make a positive difference that far out, and encouraging people to ignore present suffering may make them more callous and actually make the world worse, so EA should retain a major focus on reducing extreme poverty, animal suffering, and climate change now.

    “whether, if you encourage people to ignore present suffering, you’re actually going have a long-term drawback in that people may become more callous”

0.44

Institutional protections like tenure and platforms for controversial ideas remain necessary despite imperfect implementation.

6 pointscentrality 2/5
  • Freud was not a great philosopher because a great philosopher would be more open to different ideas, whereas Freud was highly authoritarian, fixed in his ideas, and unwilling to brook opposition—which is why Singer's grandfather, while acknowledging Freud's greater genius, broke with him and sided with Adler.

    “if he were a great philosopher, he would have been more open to different ideas. Clearly, he was highly authoritarian in his thinking.”

  • Much professional philosophy is boring because intense competition for tenured jobs and peer review push young scholars toward narrow work that critiques and builds on existing highly-rated-journal debates—aiming at 'the capillaries' rather than 'the lifeblood' of the subject—and improving this would require changing the appointment and tenure system, perhaps by giving students more of a role.

    “instead of aiming for, really, the lifeblood of the subject, it’s aiming for the capillaries of the subject”

  • There is a genuine need for the Journal of Controversial Ideas: it publishes worthwhile peer-reviewed papers that were rejected elsewhere, such as 'Merit in Science' (29 co-authors, two Nobel laureates) which argued research positions and grants were being awarded on diversity-and-inclusion criteria rather than merit and which PNAS declined as potentially harmful; the journal also allows pseudonymous publication, used by about a third of authors, especially on transgender issues where academics have faced harassment.

    “an article called “Merit in Science,” which actually had 29 co-authors. I think two of them were Nobel laureates.”

  • Tenure should not be abolished because it has protected people with controversial ideas, and recent controversies show university administrations failing to stand up to social-media outrage—suspending, dismissing, or not reappointing untenured people for controversial views—so tenure remains a needed, if insufficient, protection.

    “I worry that people with controversial ideas would simply get fired by deans when people on Twitter start criticizing.”

  • Over Singer's 24 years at Princeton, students have become harder to get to do the assigned reading—when he first arrived nearly all had read what was set, whereas now they read less—while the level of idealism among students has not changed dramatically in either direction.

    “It’s actually got harder to get them to do the reading... I think they’re reading less now, and that’s disappointing.”

  • At present the most common topic crossing the desk of the Journal of Controversial Ideas is transgender issues, where many academics publish pseudonymously because those who published work seen as 'transphobic'—often a term used too broadly—have been severely abused and harassed.

    “At the moment, I would say it’s transgender issues.”

0.43

Nature's predatory structure provides no moral justification for imitating it; we rationally combat natural suffering where able.

3 pointscentrality 3/5
  • The way nature works is a very powerful argument against the idea that the world was created by an omnipotent, omniscient, and omnibenevolent creator—such a creator seems impossible given natural suffering—yet this is the world we live in.

    “I think it’s a very powerful argument against the idea that this world was created by an omnipotent, omniscient, and omnibenevolent creator.”

  • The inference that because nature is predatory and 'red in tooth and claw' we ought to imitate it is a bad argument; the fact that nature does something is not something we ought to model but often something we should combat, just as we regret and build institutions to prevent wars even if war is part of nature, so we can rationally regret most predation while leaving it in place because we lack power to stop it.

    “the fact that nature does something is not something that we ought to imitate, but maybe, in fact, we ought to combat”

  • Reintroducing predators into an ecosystem merely because they were once part of the ecology is not a sufficient reason if doing so increases the suffering of prey animals; one should not subsidize carnivores per se, though there may be other consequentialist reasons such as effects on populations and the environment.

    “to do so just for its own sake, just because they were once here and were once part of the ecology, is not, to my view, a sufficient reason for introducing them if we know that is going to increase the suffering of some prey animals.”

0.38

Progress on animal ethics stalls because people resist thinking ethically about food choices due to social and personal costs.

2 pointscentrality 3/5
  • Singer historically held that countries pass through a stage of getting wealthier, eating more meat, and factory farming, before becoming more humane and treating animals better—so he was willing to swallow short-term negatives for the long-term outcome—but he has become less confident in that long-term improvement given little progress over the 48 years since Animal Liberation appeared.

    “I was prepared to swallow the short-term negatives for the long-term outcome. I have to admit that I’ve become less confident about the long-term outcome.”

  • The chief insight Singer drew from Hegel is that the nature of society at a given time, including its economic interests, influences people's ideas—an idea usually associated with Marx, but which Singer credits Hegel as the original source.

    “the nature of society at a time, including its economic interests, does influence people’s ideas... Marx really, I think, took that from Hegel.”