Sometimes my approach to a technical problem is "do stuff and see if it works" and other times it's "think about the problem for 1 to 3 years before making any moves" and I wish I knew how to articulate when to do which one.

Usually the "think for 3 years" thing is what I end up doing when it's a socio-technical feature with a strong possibility of causing harm if I get it wrong