You're right that they're not the same, but the author also effectively (and perhaps inadvertently) shows why not being Turing complete cannot possibly have any

You're right that they're not the same, but the author also effectively (and perhaps inadvertently) shows why not being Turing complete cannot possibly have any effect whatsoever on the difficulty of analysing what a program does (which is the implied purpose of declaring that a language isn't Turing complete). After all, suppose Java and Python changed tomorrow so that a program panics whenever a loop exceeds some large constant number of iterations. Such a change will immediately make all of programs in those languages decidable without having any effect on their behaviour. Clearly, this zero-impact change will not make those programs easier to analyse.