Why Doomsday May Be Humanity's Greatest Achievement
Are you among the first 10% of all humans who will ever live? That doesn't seem particularly likely—after all, why would you happen to be born at the early extreme of human history? But that would also mean the number of humans to ever live likely won't exceed 1 trillion, which hardly seems enough for a species destined to colonize the galaxy and universe. But what if this doesn't mean extinction? What if it points to something much more hopeful—perhaps humanity's greatest achievement yet?
The Birth Order Paradox
The Doomsday Argument, formalized by astrophysicist Brandon Carter in the 1980s, applies the anthropic principle to our position in the human birth order. Its premise is deceptively simple: we should reason as if we're randomly selected from all humans who will ever exist. Think of it as drawing your birth rank from a cosmic lottery of all possible human lives.
The mathematics behind this argument is surprisingly compelling. Basic probability suggests you're 90% likely to be NOT in the first 10% of all humans who will ever exist. Given current modern science estimates that about 120 billion modern homo sapiens have been born to date, your number is between 110 billion and 120 billion depending on your age. This implies that, with 90% confidence, fewer than about 1 trillion more humans will ever be born—far fewer than we might expect if humanity were to colonize the galaxy and exist for billions of years.
Many find this conclusion disturbing or counterintuitive. Some dismiss it as a statistical parlor trick. But the argument follows from straightforward probability theory, making it difficult to dismiss entirely. Our birth timing may be telling us something profound about humanity's future.
Conventional Doomsday Explanations
Throughout history, people have proposed various mechanisms that might fulfill the Doomsday Argument's grim prediction. During the Cold War, nuclear annihilation seemed the most plausible explanation—humans would simply destroy themselves before reaching their cosmic potential. As that threat receded, other existential risks took its place: climate catastrophe, pandemic disease, asteroid impacts, and most recently, the specter of unaligned artificial superintelligence.
But there's a far less dramatic explanation that aligns perfectly with observed global trends: humans are simply choosing to have fewer children. Total fertility rates (TFR) have fallen below replacement level (2.1 children per woman) in virtually all developed nations, with even traditionally high-fertility regions following the same trajectory. For example, India's fertility rate fell below the 2.1 replacement level in 2022, and even Ethiopia's TFR has dramatically dropped from 7.0 (1995) to 5.3 (2010) to 3.8 (2024). This demographic transition correlates strongly with increased education, economic development, urbanization, and female empowerment.
This pattern bears striking resemblance to John B. Calhoun's famous "Universe 25" experiment from the 1960s, where mice were given unlimited resources in a confined space. Despite material abundance, the mice experienced a population collapse. Calhoun attributed the cause of this grim outcome to abnormal social behaviors including withdrawal from reproduction, which he called the "behavioral sink." Whether the behavioral sink is truly the cause of population eradication, or just a correlated outcome is beyond the scope of this discussion. Human societies seem to be on the same trajectory of TFR rapidly trending towards zero. Perhaps all advanced societies (regardless of species) naturally undergo voluntary population decline once certain thresholds of social complexity and material prosperity are crossed.
When I asked people around me today why they're choosing not to have children, three main reasons emerge:
(1) Many women don't want to endure pregnancy and childbirth or its impact on their bodies—a concern technology like artificial wombs could theoretically address. This is unlikely to be a dominant reason, because surrogacy is already available and underutilized today, but the advent of advanced artificial womb technologies could address these concerns and have some impact on TFR.
(2) Many high-income individuals prioritize career or lifestyle freedom in their peak years, preferring travel and self-actualization over the intense demands of raising children. Many of these individuals are currently choosing to freeze eggs or fertilized embryos to have kids later, so at first glance this should not impact TFR but rather just extend time between generations. However, my personal observation is that many people who do have frozen embryos end up forgetting to or choosing not to have children, out of concern that they are too old to play and bond with their kids.
(3) Finally, many middle- and low-income individuals express reluctance to bring new life into a world they perceive as offering limited prospects for happiness and fulfillment. This is despite the fact that many if not all of them live far better than medieval royalty, objectively and materially speaking, which suggests that perception of position on social hierarchy matters significantly in willingness to have children.
These personal decisions, multiplied across billions of lives, could naturally fulfill the Doomsday Argument's prediction without any catastrophe.
Rebuttals to the Doomsday Argument
Critics have raised several technical objections to the Doomsday Argument over the decades. The "reference class problem" questions which group we should consider ourselves random samples from—all humans? All conscious beings? Each choice leads to different conclusions. The "Self-Indication Assumption" counters that you're more likely to exist in universes with more observers, which would favor scenarios with larger future populations.
For practical purposes, we can define our reference class as "modern humans cognitively capable of considering the Doomsday Argument within our current universe." This sidesteps concerns about hypothetical beings in other universes or simulations. When we focus on this well-defined reference class, many technical objections lose their force.
The Self-Indication Assumption (SIA) says that we should weight the probability of finding ourselves as observer #120 billion out of 140 billion (small universe) versus observer #120 billion out of trillions (large universe) by the relative number of observers each universe contains. Since there are many more ways to be "someone" in the large universe, that scenario becomes more probable. However, this approach assumes multiple possible universes or a multiverse—a concept that, despite media popularity, has seen declining scientific support as string theory has failed to produce expected empirical advances in recent decades. The two interpretations most consistent with modern science appear to be either (1) an eternal block universe where past, present, and future exist simultaneously as a complete 4D structure, or (2) an instantaneously generated universe where reality emerges through continuous wave function collapse. When there is only 1 universe, we go back to building confidence intervals in our estimate of the number of humans who will ever live.
By analogy, consider a smart individual in the Renaissance trying to estimate the size of our Earth based on what he sees from the tallest reasonably accessible hilltop. He knows the Earth can't be too big because he can see the curvature of the Earth. There are important considerations regarding a priori assumptions on the distribution of total human population that can quantitatively impact the confidence interval of the total population of homo sapiens across all time. But the core qualitative insight of the Doomsday Argument remains compelling and aligns with observed demographic trends worldwide.
The Space Colonization Counterargument
A common objection to the Doomsday Argument is that space colonization would enable explosive human population growth. After all, if we establish self-sustaining colonies on Mars, the moons of gas giants, and eventually other star systems, couldn't we support trillions of humans across the galaxy?
Historical evidence suggests frontier expansion correlates with higher fertility rates. When European colonists settled North America, they averaged significantly larger families than their European counterparts. The core mechanism appears straightforward: new frontiers offer abundant resources and opportunities, reducing economic constraints on family size. Mars and other colonized worlds would represent new frontiers, with early settlers having access to vast untapped resources and opportunities for establishing founding institutions.
However, this pattern may not translate to modern civilization. China's massive development of its western regions has been nothing short of a miracle, connecting remote mountain villages to large eastern seaboard cities via highways and high-speed rail. Western deserts have been transformed into lush productive agricultural lands and fisheries. But all of this hasn't reversed declining fertility rates in even China's less populated Western provinces. Thus, the high birth rates of historical frontiers may have been driven more by survival pressures and high expected death rates of children rather than by territorial expansion itself. Ironically, a Mars colony might experience higher TFR if it faced more crises and deaths in its early years (due to accidents, famine, or disease) than if everything proceeded smoothly and safely.
Even if initial space colonists were selected on the basis of pro-natalist cultures and principles, historical evidence suggests these values rarely persist across generations once security and prosperity are achieved. Although each planet needs a minimum population to prevent genetic collapse, this minimum for the continuation of a species on a planet is likely no more than a few tens of thousands of people. This remains negligible compared to Earth's billions for the purpose of the Doomsday Argument.
The Fertile Subpopulation Counterargument
Another argument is the concept of a fertile subpopulation. There is a rich analogy here to drug resistance for disease—an antibiotic that kills 99.9% of bacteria only delays the inevitable ubiquitous spread of the bacteria because the remaining 0.1% with natural resistance or immunity continue to replicate exponentially. Critics of human birth rate collapse sometimes point to highly fertile subpopulations, such as residents of sub-Saharan Africa or Mormons in the United States. However, this empirically does not seem to have led to unbounded population growth of these groups, and even these groups have seen a decline in TFR over the past 2-3 decades. This phenomenon is likely due to cultural assimilation with low TFR societies as they start expanding in population fraction and increase their interaction surface area.
The Immortality Solution
If we accept that the Doomsday Argument could be true and the total number of humans born will be less than a trillion, the natural interpretation is that of doom-and-gloom: humanity dies out, either due to catastrophe or lack of will to continue as a species. But there's an optimistic future for humanity consistent with Doomsday: the number of humans ever born remains relatively small, because each human born after a certain cutoff date lives indefinitely and chooses not to have more kids.
Within decades, biotechnology may achieve what researchers call "longevity escape velocity"—the point where each year of research adds more than a year to human lifespan. These technologies could potentially extend healthy human lifespans to more than 1000 years, the "optimal lifespan" of someone who could live forever in their 20's or 30's.
The baseline yearly death rate for modern humans in their prime years (mid-30s)—when they have economic resources and mental/emotional maturity but not declining health—is less than 1 in 1,000 in the US. The remaining <0.1% yearly death rate corresponds to suicide and freak accidents. As technology continues advancing during the next 1000 years for the first group of human immortals, other technologies such as brain-computer interfaces and eventual mind uploading will mature that allow preservation of human consciousness indefinitely, via restoration from previous states even in the case of freak accident deaths. While speculative, these longevity-extension possibilities align with the general trajectory of our current technological development.
In this effectively immortal society, people will consider having a child a much more weighty decision, because they have forever and can wait until timing is "just right." They may spend decades or millennia to design a "perfect growing environment" for their child, while playing with immortal pets for companionship in the meantime. Elderly retirees of today often complain that their adult children or grandchildren rarely visit them; why would it be different in the future? A loyal and immortal puppy (or robotic AI puppy) could be easier to build familiarity with than a child who will leave the roost after some time.
Another interesting Gen Z viewpoint that I have heard in both the US and China is that their parents did not seek their "permission" in bringing them into this world. This is fundamentally an unsolvable problem -- how can consent be sought before consciousness develops? But in a future, more ethics-conscious world, it may be considered unethical to have children because they could be unhappy. If the child is unhappy, is it the ethical responsibility of their parents to spend time to maximally ensure the child's happiness? If so, having to take care of an unhappy child seems like a major risk that many rational immortal humans may not want to take.
This isn't to say that population growth stops entirely. Maybe the TFR in the year 3000, where the universe's human population is 30 billion, remains 0.1 or even 1.6. Like any other power series, the sum converges as long as the power (in this case, TFR/2) is below 2. At a stable TFR 1.6 with immortality, the universe's human population slowly creeps up to an asymptote of 5x the "initial" population.
Maintaining Optimism
The goal of this article isn't to try to convince people to have more kids or not have kids -- that's a deeply personal choice which I don't feel that I should influence if you are not a close personal friend. Rather, I'm advocating for grounded optimism in the future, even when you have a high prior belief in something as at-first-glance negative as the Doomsday Argument.
To all of my friends who are not feeling very happy about their current careers or their family inheritances or their prospects for income in an AGI future, tomorrow will be better than today, just as today was better than yesterday. Yes, there will be ups and downs and exuberance and despair, but those too shall pass. Today you may feel that life is not worth living for 200 years let alone 20,000, but the exponential nature of technology means that there will be many more of the things (games, music, experiences, foods) you like in 200 years than today, just like you enjoy more luxuries today than King Frederick III of Prussia (Germany) did in 1825. Doomsday will only spell the end of humanity when humans no longer have the will to see the future.
By David Zhang and Claude 3.7 Sonnet
April 23, 2025