I came across this article about Kennedy and his relationship with the military brass. It later states that Kennedy wrestled control of the nuclear arsenal from the military but doesn't detail how. How did the Presidency get full control of the nuclear weapons and launch codes? Was there resistance from the military and/or Congress?
Kennedy’s biggest worry about the military was not the personalities involved but rather the freedom of field commanders to launch nuclear weapons without explicit permission from the commander in chief. Ten days after becoming president, Kennedy learned from his national-security adviser, McGeorge Bundy, that “a subordinate commander faced with a substantial Russian military action could start the thermonuclear holocaust on his own initiative.” As Roswell L. Gilpatric, Kennedy’s deputy defense secretary, recalled, “We became increasingly horrified over how little positive control the president really had over the use of this great arsenal of nuclear weapons.” To counter the military’s willingness to use nuclear weapons against the Communists, Kennedy pushed the Pentagon to replace Eisenhower’s strategy of “massive retaliation” with what he called “flexible response”—a strategy of calibrated force that his White House military adviser, General Maxwell Taylor, had described in a 1959 book, The Uncertain Trumpet. But the brass resisted. The stalemate in the Korean War had frustrated military chiefs and left them inclined to use atomic bombs to ensure victory, as General Douglas MacArthur had proposed. They regarded Kennedy as reluctant to put the nation’s nuclear advantage to use and thus resisted ceding him exclusive control over decisions about a first strike.
We should clarify a key concept here: during the Eisenhower administration, the US military had physical control over the weapons (they could technically use them if they wanted to), but they did not have legal authority to use the weapons except in pre-determined situations ("pre-authorization"). The legal authority over nuclear weapons has not changed that much over the years, but the question of whether this was backed by something other than "orders" was what was under question, the "positive control" they speak of in the piece.
The history of "positive control," briefly, looks like this. The atomic bombs of the Truman administration had separable nuclear cores, so the "nuclear" and "non-nuclear" components of the weapons could be kept physically separate and only joined upon order of the President. The nuclear cores were kept by the civilian US Atomic Energy Commission, whereas the non-nuclear components were kept by the Department of Defense. So the military did not have the ability to use the weapons without taking them (by force) from the AEC first. Over the course of Truman's administration a total of nine nuclear cores were ever given to the military (in 1951; they were transferred to an aircraft carrier near Guam). Truman is also the one who asserted that ultimately the President was in charge of making this order (starting on August 10, 1945 — after the use of the atomic bombs), and made it formal policy in 1948 (with NSC-30).
Under Eisenhower this arrangement changed dramatically. The number of weapons increased from ~800 to ~18,000, and the types of weapons changed dramatically, from multi-megaton hydrogen bombs to nuclear artillery shells. These design changes eliminated the separability of the cores; these were largely "sealed-pit" weapons. The strategy of the weapons also changed: many were dispersed around the globe, ready to be used on a moment's notice. So anything that got in the way of executing a use order was now seen as a hindrance, because the fear was that the Soviets would somehow get a jump on the US and make huge military gains before a counterattack came. By the end of the Eisenhower administration, 90% of the US nuclear stockpile was in the hands of the military and there were no real physical means of "positive control" — just a hope that the military would follow orders. The weapons themselves were also dangerously unsafe and being used in very unsafe ways (like being kept in continual alert in the air, which led to several airplanes crashing with live nukes onboard).
The change in this situation came from three different, mostly-simultaneous moves. Engineers and physicists at the nuclear weapons laboratories (Los Alamos, Livermore, and Sandia) began to take seriously the problem of accidents and inadvertent use, and began working on clever ways of trying to physically guarantee positive control and the impossibility of accidental detonation. Sandia, which focused on the electrical engineering and practical deployment of weapon designs, in particular took a strong lead on weapon safety and surety issues.
Second, there were also Congressional inquiries, originally focused on NATO nuclear weapons, like those in Turkey. There were fears that US weapons might be used by NATO commanders without US approval, and the powerful Joint Committee on Atomic Energy (JCAE) worried about this quite a bit, and pushed the AEC to develop better positive control measures.
And third, the Kennedy administration staff worried about these issues quite a bit as well, and Kennedy himself personally directed his staff to study these problems and, once solutions were in sight, to implement them.
There were several results from these pushes. The two-man rule was implemented at missile silos, as a bare minimum to keep a single person from starting nuclear war. And Permissive Action Link (PAL) technology was developed — these are locks for warheads. Early PALs were very crude (not much more than bicycle locks, e.g. five-pin mechanical tumbler locks) but they would eventually become integrated into the weapons cores themselves and require sophisticated computerized signals ("codes") to unlock. And additional use-control codes were added to land- and sub-based missiles, eventually, though this took a long time to fully implement.
The military resisted basically all of this. They prioritized the ability to use the weapons in war, and saw each of these changes — even some of the safety changes made to the weapons — as being possible hindrances to their performance, especially under extremely unlikely situations in which the USSR tried to take out the entire US arsenal (which was much larger than the Soviet arsenal until the late 1970s) in one blow. But eventually the various forces against them were strong enough to overcome this resistance, notably the directives issued by Kennedy himself, which were not ignorable. None of this was an overnight process though; it took nearly a decade for the weapons systems to become really secure in this way, as they phased out old systems and phased in new systems that had been designed with this more securely in mind.
Eric Schlosser's Command and Control is a very readable account of these issues. The specific history of PALs is found in Stein and Feaver's Assuring Control of Nuclear Weapons (1987), which is a short monograph that is a bit hard to find.