Abstract
Positively influencing the long-term future should be a key moral priority of our time, because future people matter no less than present ones, there could be astronomically many of them, and this century’s decisions — over existential risks and value lock-in — may determine whether they exist at all.
Framing
- Opens with Derek Parfit’s buried-glass case: a child cut by broken glass in 5 years and one cut in 500 years are harmed equally, so there is no reason to stop caring about effects merely because they are distant in time.
- Concern for future generations is already commonplace (children, grandchildren); longtermism extends it without arbitrary cutoff.
- If human history were a novel, we may be on its first page.
- Two-sided claim about the present moment: uniquely rapid change creates both opportunity (values embedded in new technologies could benefit everyone long into the future) and a possible “time of perils” — a period in which catastrophe severe enough to permanently curtail humanity’s potential is achievable.
Pillar 1 — Future people matter
- Spatial analogy: learning that a friend in need has moved to another continent does not reduce how much their situation matters, provided helping is equally easy.
- If birthplace does not affect moral status, birth date should not either. From our ancestors’ perspective, we were the future people, and our joys and pains are no less real than theirs.
- Concedes there can be derivative reasons to weight nearby people more — kinship, or greater tractability of helping them — but never the bare fact of closeness in space or time.
Pillar 2 — Humanity’s potential is enormous
- Analogy from individuals to the species: part of what makes saving a life valuable is preserving that person’s potential; humanity likewise has a latent potential that can be protected or foreclosed.
- Species baseline: a typical mammalian species lasts around a million years; Homo sapiens has lasted ~300,000, implying at least ~200,000 years and ~1,000 generations remaining — and humans have technological means of preventing and recovering from threats that other mammals lack.
- Planetary baseline: Earth stays habitable for several hundred million years before sterilisation by the sun. Surviving just 1% of that gives ~500,000 further generations, or at least a billion billion future lives — ten thousand times all humans to date.
- Beyond Earth: only 66 years separated the first heavier-than-air flight from the Moon landing; spacefaring over centuries or millennia looks possible, making the future potentially vast in extent as well as duration.
- Jeans’ scale metaphor: a postage stamp = recorded civilisation; stamp plus coin = human species lifetime; a 20-metre obelisk under the coin = age of the Earth; a 5-metre tree atop it = Earth’s habitable future; the Matterhorn behind it = the habitable future of the universe.
Why that future could be good
- Historical progress is offered as evidence: extreme poverty fell from ~90% of people in 1820 to under 10% in 2015; child mortality from over 40% to under 5%; the share living in a democracy from under 1% to most of the world.
- Appreciating how good the future could be is framed as extra motivation for present problems, not a substitute — solving poverty, injustice, animal suffering, and climate damage now could come close to solving them permanently.
- The symmetric case: the future could instead be stagnant, totalitarian, or conflict-ridden. This strengthens rather than weakens the argument, since preventing future tragedy matters as much as enabling future flourishing.
Pillar 3 — Our actions could influence the long-term future
- For most of history the honest answer may have been that nothing could be done; the claim is that this moment is unusually influential.
Climate change as proof of concept
- Human-caused climate disruption is established, and some effects persist for tens of thousands of years because atmospheric CO2 is long-lived.
- The amount of damage is under our control — green technology, zero-carbon energy, carbon pricing at true social cost — which demonstrates that long-run influence is possible.
Existential risks
- Nuclear weapons: since the Trinity test in July 1945, arsenals grew into the tens of thousands, many still on hair-trigger alert; large-scale war could cause a prolonged nuclear winter and worldwide crop failure.
- Engineered pandemics: biotechnology may allow pathogens far deadlier or more transmissible than natural ones. Barriers are falling — the Human Genome Project took ~15 years and half a billion dollars to 2003, while a genome can now be sequenced in about an hour for ~$1,000, and DNA synthesis costs have fallen by a factor of over 1,000.
- Artificial intelligence: experts including Stuart Russell warn that general-purpose, above-human-level AI is plausible within our or our children’s lifetimes. The primate analogy — human dominance rests on intelligence, not strength or intent — motivates the concern that systems with the wrong values could effectively end human control over the future. Alignment is presented as possibly the most important challenge of the century.
- These three are given as illustrations rather than a ranking: a century ago essentially nobody could conceive of them, and the same may be true of risks ahead. A highly networked global society may be unusually vulnerable.
- Toby Ord (The Precipice) is quoted: risking the future for present advantage “privileges a tiny sliver of our story over the grand sweep of the whole.”
Institutions and value lock-in
- Political reform to represent future generations: national committees and offices, new voting methods such as Demeny voting, international frameworks and panels; the UN secretary-general’s Common Agenda is cited as movement in this direction.
- Periods of rapid change tend to fix contingent moral and political choices for long stretches. The present qualifies — technological advance, globalisation, and historically meteoric economic growth.
- A further intervention follows: making it less likely that any single set of moral, cultural, or political values dominates prematurely, before all perspectives can be heard.
Putting it together
- The moral premise is widely shared and easy to accept precisely because it seems to carry no practical implications; the essay’s work is done by the empirical claim that our actions do reach that far.
- Concrete examples of what this implies: pandemic countermeasures such as early detection systems for novel pathogens; advocacy for institutions protecting future generations; research in economics, history, law, or philosophy on long-run influence mechanisms; technical work making powerful AI systems transparent and aligned.
Conclusion
- Anticipates the objection that the ideas sound like science fiction, and responds that once-avant-garde moral ideas (humane treatment of animals) are now commonplace; strangeness is not disconfirmation.
- Longtermism is presented as a broad perspective tying together many research programmes, not a single narrow prescription — partial agreement is enough to find it useful.
- Uncertainty is acknowledged, but claiming that nothing can be done for the long run is called extremely overconfident; the response to uncertainty is further inquiry rather than resignation.