Worse is when you find those dead ends while working for somebody who doesn't understand software. I've seen consultants stiffed for weeks of billable hours for researching technologies they were specifically asked to research.

"In the end, it was faster to roll our own than integrate X, and we'll be able to move quicker in the future."

"Then why did you waste all that time playing with X???"

On the other hand you needed those 2 dead ends in order to establish that one way, and only one way, is the right one.

You cannot expect yourself to be that foresightet.

Not just that. You also end up with the knowledge that those 2 paths were actually dead-ends. It's important to know solutions that will NOT work.

One benefit of this kind of work is that the two dead ends don't end up as pieces of code you have to maintain thereafter.