Was thinking yesterday, I'm starting to get more bothered by the phrase "Well it's a hard problem to solve" and leaving it there. 

Papering over complexities with more tools often defers responsibility to somewhere else in a stack and often forces users to trust the system and it's designers decisions. 

Maybe it's okay to toss things in the bin and start over sometimes, reducing complexity for users by reducing complexity in the system. I think loss aversion also plays a part, maybe some features need to be sacrificed for the sake of longevity. 

Much smarter people than myself have pondered these things, but they often lead me to question their bias toward their sunken cost and risk exposure to the success of the system, clouding their ability to hear a reasonable doubt.