It’s the End of the World As We Know It…But Not the End of the World
August 31, 2026
By Matthew Pietz
This article was written and edited without the use of AI.
In a typical month you’re probably in several conversations about AI or robots; and probably in at least one of them, someone (maybe you) makes a joke-not-a-joke that we’re living in end times.
There is a palpable sense of apocalypse in the air these days. Books by “doomers” like Eliezer Yudkowsky top best-seller lists; Hollywood is increasing its output end-of-the-world stories to meet public demand; and tech-related anxiety is spiking.
But you may find some comfort in knowing this isn't the first time people have felt this way. Several generations before us have been totally convinced the end of the world was imminent. So far… well, they were totally wrong.
Aren't things different this time? Yes, but at Keranut we feel they are not different in ways that mean the apocalypse is now, or anytime soon.
For a tour of apocalyptic belief over the millennia, we recommend Dorian Lynskey’s 2024 "Everything Must Go". He traces phenomena like the Second Great Awakening in 1843, when public belief that God was going to end the world was so strong that tens of thousands of Americans gave away their money, left crops to rot in the field, closed businesses and turned loose livestock to run wild.
More recent apocalypse concerns were more based in science, but also wrong in their absolute predictions. The “population bomb” theory of the 1960s, which inspired China’s one-child policy, led to a paper predicting that in November 2026—three months from now—the human population would approach infinity. Population growth has certainly driven environmental damage and climate change, but the predicted mass starvation never happened, sixty countries have declining populations, and the global population is projected to decrease starting as early as the 2060s.
Fear of nuclear annihilation, present in the popular consciousness right from the advent of the bomb in 1945, has ebbed since its peak in the 1980s, as proliferation slowed and nuclear states developed safeguards for accidental or precipitous launch. There have been terrifyingly close calls, and we are not out of danger, especially as US-Russia arms control treaties have been dissolved. But most analysts feel the full-scale war that would lead to nuclear winter, killing just about everyone, is highly unlikely.
Climate change was the next source of apocalyptic belief, arriving in the 1980s, and here, again, we are certainly in peril. Fires, floods, heat waves, crop failure, water scarcity, disease and economic damage are already here, and going to get much worse. But the apocalypse scenario of 8.5° C increase by 2100, hugely influential and the default of thousands of studies and projections,, has now been discarded by the scientists who first came up with it. Some will die, and the rest will adapt.
This, we feel at Keranaut, provides a good model for what we can expect from the lightning-speed advancement of AI and robots. Government efforts to introduce safeguards have so far been ineffectual (though China does vastly better than the US), as have the occasional calls from one Silicon Valley figure or another for a voluntary slowdown. The behemoth of advanced tech will continue to accelerate, and it will wreak some havoc along the weay.
There will be economic fallout, and there may also be more visceral consequences. Attacks on populations via bioengineered viruses or fleets of tiny, cheap, and coordinated drones; rogue AIs bringing down public services; military autonomous robots firing on civilian targets—all of these threats are credible, and in the worst case scenarios, millions of lives could be lost.
But such tragic events may also give global leaders the courage to stand up to a tech sector that lobbies relentlessly to be left alone, and finally implement meaningful safeguards. It may sound grim to say we need a major disaster before we can be safe, but it is much more likely that we will have smaller, non-world-ending events that we can adapt to, than that one new thing will take the world by total surprise and wipe out the population in one fell swoop. The Cuban Missile Crisis, Chernobyl, and many lesser-known incidents were arguably the prime drivers of safety measures that helped contain the nuclear threat. Catastrophe followed by reform is hardly a model to aspire to. But it is a more likely path than the predictions of the Doomers.
Don’t take our word for it. MIT recently asked 272 experts to rate the highest risks from AI, and the resulting analysis identified 5 threats that have a 12%-22% chance of happening in a way that causes more than one million deaths (if no preventative steps are taken from now). Those are sobering numbers; but none of the projected events would come close to causing the end of humanity, or the planet.
In “Everything Must Go”, Lynskey argues that some apocalyptic thinking comes from the human need to feel important, that seeing oneself as a tragic actor in humanity’s final chapter feeds the ego. Let’s have the humility to know we’re going to survive.
Things will radically change in the next decade, in many good ways, and with terrible things happening along way. But just as with climate change, there is no reason for the despondency of doomerism: because we are going to survive, we should act, as we still have time to do, to ensure we survive as well as we possibly can.
Click here to subscribe and be notified of future posts.
Cover photo by Rajesh Rajput on Unsplash
