Why should the game end at all? Why shouldn't people continue developing better and better shells that people can use to interact with their computer, or maybe different ones for different use cases? Supply-chain attacks are bad and worth mitigating for any kind of software, not just your shell; but the possibility of such an attack doesn't mean it's inherently unwise to try out new pieces of software. Saving time on your shell history search is good and declaring it unimportant merely because the amount of time saved sounds small, is how we wind up after many iterations with software that is noticeably laggy to the end user. But the real value of new shells I think is the new features you didn't know you would find useful at the time.