On September 12, 1933, the Hungarian physicist Leo Szilard was waiting for the traffic light to change on Southampton Row in Bloomsbury when he worked out how to make a nuclear chain reaction. He was 35 and living in London as a refugee, having left Germany in the spring.
What provoked him was a remark by Ernest Rutherford in The Times that morning. Rutherford had called it moonshine. Szilard took it as a claim about physics, not engineering. It annoyed him. He went for a walk. The reasoning he arrived at is short. If an element could be found that released two neutrons for every one absorbed, and if enough of those neutrons were captured by other nuclei rather than escaping, then each generation would produce more than the last. The reaction would run by itself. The energy released would be enormous, and so would the military implications.
There was no such element known at the time, and Szilard did not claim one. He filed a British patent application in 1934 covering the chain reaction and the idea of a critical mass, and he assigned it to the Admiralty so that it would be kept secret. Neutron-induced fission was not discovered until December 1938, by Hahn and Strassmann in Berlin, and measured and explained by Meitner and Frisch in January 1939, and when it was, Szilard's 1933 reasoning applied to it directly.
He spent the next decade on the politics of it. He and Eugene Wigner drafted the letter signed by Einstein and sent to President Roosevelt on August 2, 1939, warning that Germany might build a bomb and urging the United States to start; he was one of the first scientists at Chicago and worked with Enrico Fermi on the first controlled chain reaction, which ran on December 2, 1942, in a squash court under the stands at Stagg Field. He and Fermi filed the reactor patent in 1944.
Szilard then campaigned against using the bomb on cities, and circulated the petition among Manhattan Project scientists in July 1945 that reached Truman after Hiroshima. He spent the rest of his career on arms control and on molecular biology, and died in 1964.