It's funny, I have had similar experiences, with a totally different conclusion. I've worked with people who just throw shit at the wall like that, and like you, I've seen the shit stick to the wall often times. But my conclusions have been something like: 1. It doesn't work out just as often as it does, and it's impossible to predict which situation is which, and harder to fix post-facto than it would have been to just not do it that way from the start, so (IMO) it's a crapshoot with negative expected value, 2. I hate working this way, at a personal level, which leads to: 3. I just don't want to spend my time working on stuff where this is even plausibly the right approach.

All in all, this was one of the formative experiences that has led me to seeking out deeper and larger projects to spend my time on, rather than the string of smaller more trivial things that lend themselves to this approach to software. But the other side of that coin is that I'm a lot more content now with the understanding that other people just have different preferences than me. I spend a lot less time thinking about what the right way to build software is, and more time thinking about what's right for me.