On 7 June 1742, a Prussian mathematician named Christian Goldbach sent a letter to Leonhard Euler with a simple idea about prime numbers. Almost three centuries later, computers have checked it for every even number up to 4 quintillion, and nobody has managed to prove it.
The idea is so easy to state that a child can test it. Take any even number bigger than 2. You can always split it into two prime numbers added together. 4 is 2 plus 2. 10 is 3 plus 7. 100 is 47 plus 53. It has never failed. And yet it remains one of the oldest unsolved problems in all of mathematics.
A letter, a reply, and a problem that refused to die
Goldbach did not publish his idea in a grand treatise. He wrote it in a private letter to Euler.
Euler answered on 30 June 1742, barely three weeks later. In his reply, he reshaped Goldbach’s original proposal into the clean “strong” version still quoted today: every even number greater than 2 is the sum of two primes.
Euler was convinced. He described it as “a completely certain theorem”. But even Euler, a man who proved results in almost every corner of mathematics, could not find a proof.
The idea was not really Goldbach’s first
Here is the twist most people never hear. Before Goldbach ever wrote his letter, René Descartes had already suggested something very close: that any even number could be written as the sum of at most three primes. A slightly weaker claim, but clearly the same family of idea.
So why does the problem carry Goldbach’s name? The legendary mathematician Paul Erdős had a famously blunt answer. He admitted that “Descartes actually discovered this before Goldbach… but it is better that the conjecture was named for Goldbach because, mathematically speaking, Descartes was infinitely rich and Goldbach was very poor.”
Checking it by hand, then by machine
For centuries, the only way to gain confidence was to test numbers one by one. In 1938, the mathematician Nils Pipping did exactly that, checking the conjecture by hand all the way up to 100,000. No computer, no calculator, just patience and paper.
The computer age changed the scale completely. The researcher Tomas Oliveira e Silva led a distributed computing search that pushed verification up to 4 quintillion, a 4 followed by 18 zeros. Along the way, the search turned up some stubborn cases where the only way to make the sum work needed surprisingly large primes.
Still, no matter how many numbers pass the test, a check is not a proof. There are infinitely many even numbers, and a single exception hiding beyond the last one tested would be enough to sink the whole idea.
Mathematicians keep getting closer without arriving
Progress has come in partial victories. Between 1937 and 1938, Chudakov, van der Corput and Estermann showed that almost all even numbers can be written as the sum of two primes. Almost all, but not provably every single one.
In 1973, Chen Jingrun went further. He proved that every large enough even number is either the sum of two primes, or the sum of a prime and a number made of exactly two prime factors. Painfully close to the real thing.
Then, in 2013, Harald Helfgott proved the “weak” Goldbach conjecture: every odd number greater than 7 is the sum of three odd primes. A major milestone, and it has a neat consequence. It means every even number from 4 upward is the sum of at most four primes. The strong version, though, still stands untouched.
A tiny machine that could settle everything
One of the strangest facts about this problem comes from computer science. There exists a theoretical machine, a Turing machine with only 27 states, that is designed to stop running if and only if Goldbach’s conjecture is false.
Put simply, it hunts forever for an even number that breaks the rule. If it ever halts, the conjecture is dead. If it runs forever, the conjecture is true. The catch is that nobody can predict what this little machine will do, which ties a 1742 letter directly to the deepest questions about what computers can and cannot know.
Why such a simple question still matters
Goldbach’s conjecture is a reminder that the hardest problems in mathematics are not always the ones filled with complicated symbols. Sometimes they are sentences a schoolchild can understand, backed by quintillions of successful checks, and still waiting for someone to explain why they must always be true.
Somewhere out there, the proof may already be hiding in plain sight. Nearly 300 years after a friendly exchange of letters, the question is who will finally find it.

