Many things work in theory.
There's a time for thinking in generalities, and a time for testing our projects against the world. They are not the same time.
As a rule: I aim to only do things that are hyper-general (like these principles) or hyper-specific.
It should be clearly obvious whether I've succeeded or failed, based on market feedback, the laws of physics, or both.
Of course, this can be quite painful.
When we work hard on something, it hurts to see it die.
But consider the alternative: with no criteria for failure, it's impossible to know if we've succeeded.
This can be overdone. "No risk, no reward" can be used as a socially acceptable explanation for bad decisions.
On the other hand, without failing specifically, it's hard to learn specifically. And without that, we're flying blind.