Puzzles (1500-1600)

I keep seeing a lot of puzzles that are specifically designed to make a play around the opponent playing a blunder. I really don't understand this, because it's trying to reinforce that I should see these ridiculous end states because I took a move that would give me an even larger lead IF they made that blunder as opposed to making a strong play.

Don't get me wrong, not all the puzzles are like this, but I'm often left with more questions than answers when the computer blunders their queen, or dives a bishop into a pawn line for no reason. Why would anyone do that, ever? Even looking into the lines the engine will be like "yup that's a total 100% loss move with 0 recovery", so I don't see why it's reinforcing this kind of logic at all.