Are we living at the end of human history?
If your birth rank is randomly sampled from all humans who will ever live...
The argument says you're likely in the middle 95% of all humans, not the first 2.5% or last 2.5%.
Imagine two urns in front of you. One contains 10 balls, the other contains 1 million balls. Each ball is numbered. You don't know which urn is which.
You reach in and pull out ball #7.
Ball #7 is in the 70th percentile. Very plausible!
Ball #7 is in the 0.0007th percentile. Extremely unlikely!
The Doomsday Argument applies this reasoning to humanity. Your birth rank (~95 billion) is your "ball number." If humanity will produce trillions of future humans, your existence at #95 billion would be extremely improbable—like drawing ball #7 from the million-ball urn.
Humanity goes extinct within the next few centuries
total humans ever
You're in the 47th percentile—perfectly reasonable!
Humanity colonizes the galaxy and thrives for millions of years
total humans ever
You're in the 0.05th percentile—incredibly lucky!
The Doomsday Argument doesn't prove we're doomed—it says we should update our probabilities. If you thought there was a 50/50 chance between these scenarios, the argument says you should now think "Doom Early" is about 1000× more likely.
The Doomsday Argument was first formulated by astrophysicist Brandon Carter in 1983, though he didn't publish it. It was independently discovered by physicist J. Richard Gott and philosopher John Leslie, who brought it to wider attention.
Philosopher Nick Bostrom has extensively analyzed the argument in his book Anthropic Bias: Observation Selection Effects in Science and Philosophy, exploring both its implications and potential flaws.
Who counts as "human"? Should we include only Homo sapiens? What about future posthumans or AI? The choice of reference class dramatically changes the prediction, yet there's no principled way to choose.
Dennis Dieks argues that your mere existence provides evidence that more observers exist. Under SIA, the Doomsday Argument's conclusion is exactly cancelled out—you can't infer anything about total population from your birth rank.
Statisticians Gelman and Robert note: "While 95% of confidence intervals will contain the true value of N, this is not the same as N being contained in the confidence interval with 95% probability." The Doomsday Argument may conflate frequentist and Bayesian probabilities.
Here's what makes this a true paradox: The argument uses seemingly reasonable probabilistic reasoning (we shouldn't be "special" in our position among all humans) to reach a startling conclusion (humanity likely ends soon). Yet something feels deeply wrong about predicting doom from pure statistics.
Either: (1) the argument is valid and we should genuinely update our extinction estimates upward, (2) there's a subtle flaw in the probabilistic reasoning that philosophers haven't yet identified, or (3) anthropic reasoning itself is more mysterious than we thought.
After 40 years of debate, there's still no consensus. The Doomsday Argument remains one of the most controversial applications of probability theory to existential questions.