Jeremy Bernstein in the New York Review of Books blog:
It should be noted that when nuclear weapons first began to be constructed in the early 1940s, no one thought of deterrence. The bomb was not designed to “deter” Hitler. It was to defeat him and his Axis allies. In the spring of 1943 the Columbia physicist Robert Serber gave a series of lectures to new recruits at Los Alamos. The opening lines of the printed version read: “The object of the project is to produce a practical military weapon in the form of a bomb in which energy is released by a fast neutron chain reaction in one of more of the materials known to show nuclear fission.”
As far as I can tell, the first suggestion that these weapons could be used for deterrence came from General Leslie Groves, who headed the Manhattan project. Some time after the defeat of Germany, but well before the first successful test of the bomb in July of 1945, he came to Los Alamos. At a small dinner he expressed the view that the Russians would have to be deterred by the bomb. He was sure that they had expansionist plans that included the domination of all of Europe and that nuclear weapons would be necessary to stop them. In fact, Soviet spies had already furnished Stalin with extensive knowledge of the US program well before it became public after Hiroshima; Stalin’s reaction was not to be deterred, but to start a crash program to build nuclear weapons of his own while at the same time occupying the countries of Eastern Europe.
Indeed, if you think about it, deterrence is an odd concept. It implies explicit or implicit negotiations between the deterrer and the deterree. How is one to know when deterrence has been successful? It is easier to know that it has not been when one is attacked. David tried to deter Goliath by invoking the God of the Israelites. Goliath had no interest. If David had shown Goliath his skills with a slingshot instead of attempting to deter him it would probably have provoked a better defense against sling shots. What does one expect a deterree to do, sign a document admitting that he/she is deterred? Would anyone trust such a document? Without such a document how much deterrence is enough? Is one atomic bomb enough? How about fifty or five hundred? Who is to decide? For many decades, the US and Russia were engaged in a policy of “mutually assured destruction”—MAD. How did we know that destruction of the other side was “assured?”