A question circulating on social media in recent weeks looks deceptively simple. Everyone on Earth casts a private vote by pressing either a red button or a blue button. If more than fifty per cent of people press blue, everyone survives. If fewer than fifty per cent press blue, only those who pressed red survive. Which would you press?
The version that went viral was posted by the writer Tim Urban on X in April 2026. Within days it had drawn more than 22 million views and roughly 95,000 votes, breaking down at about 57.9 per cent for blue and 42.1 per cent for red. The argument that followed quickly collapsed into a moral shouting match. Blue voters accused red voters of selfishness. Red voters accused blue voters of naive sentimentalism. Each side concluded the other was either stupid or wicked.
A calmer look at the structure reveals something different. This is not a moral dilemma. It is a coordination problem.
Consider the bare arithmetic of self-interest. Anyone who presses red survives no matter what happens. If blue wins more than half the vote, everyone lives and the red voter lives along with them. If blue falls short, only red voters survive and the red voter lives because they pressed red. In game-theoretic terms, red is a dominant strategy and a Nash equilibrium. A person concerned only with staying alive has no rational reason to choose anything else.
Blue is the opposite. Pressing blue stakes your life on a single proposition: that enough strangers will press blue too. If the bet wins, everyone lives. If the bet loses, you are the one who dies. The safety of blue depends entirely on your trust in people you cannot see, cannot speak to, and cannot influence.
The strange feature of the dilemma is its death curve. If every single person on Earth presses red, the share of blue votes is zero, the threshold is not met, and only red voters live. Since everyone pressed red, no one dies. If every person presses blue, the threshold is comfortably cleared and again no one dies. Both extremes, total selfishness and total altruism, produce identical outcomes: no casualties.
The danger lives in the middle. The worst case is something like 49 per cent pressing blue and 51 per cent pressing red. Blue falls just short, and every one of those forty-nine in a hundred kind-hearted voters dies. The more people choose the moral option, so long as they remain below the line, the higher the body count climbs. Half-hearted altruism kills more people than wholehearted selfishness.
This is the real point of the puzzle. The rules concentrate every ounce of risk on the cooperators while leaving the self-preserving entirely untouched. The question is not whether you are kind. It is whether you are willing to hand your life to a crowd of strangers you can neither identify nor coordinate with. Red voters are not necessarily cold-blooded. They simply do not believe coordination will succeed. Blue voters are not necessarily noble. They are willing to gamble on human nature.
The shape of this problem is everywhere in the real world. A bank run works the same way. As long as you believe other depositors will leave their money in the bank, you have no reason to withdraw yours. The moment you suspect they will run for the exit, the rational move is to beat them to it. Herd immunity, climate negotiations, the upkeep of public goods all share the structure of the red and blue button. Cooperation produces the best collective result, but the first cooperators carry the largest individual risk.
The question worth asking, then, is not whether the people pressing buttons have a conscience. It is how the rules of the game have been written. A system that piles all the risk onto those willing to help will not grow cooperation. It will breed quiet, intelligent self-preservation. What deserves serious thought is not which colour you would press, but whether it is possible to design rules under which pressing blue no longer requires courage.
