4 ideas
In the early 1960s Watson Jr. approved an estimated $5 billion program to replace IBM's own incompatible, profitable computer lines with a single compatible family, announced in April 1964. The bet made IBM's existing products obsolete, and software delays in OS/360 nearly broke the company. Watson admits he committed to the program without being sure it would work.
Large organizations naturally domesticate their abrasive, independent people. A leader must deliberately shelter the dissenters who tell him he is wrong, because once that capacity is bred out it cannot be restored.
Watson Jr. drifted as a mediocre student and a salesman carrying his father's name. He found real self-confidence only as a WWII pilot, where his competence owed nothing to his father. An heir to a dominant founder can only claim authority if he has first proven himself somewhere the founder's influence cannot explain his success. Even then, the transfer of power comes through bruising conflict rather than blessing.
Watson Sr. built IBM on punched-card tabulators and doubted that electronic computers could be a large market. His son had to push the company into electronics over his father's resistance. The mechanism is that a leader's mastery of the current technology makes the next technology look like a niche. The push to change therefore usually has to come from someone less invested in the old success.