That's in effect the question from Amanuel, who writes (via email; lightly edited):
Hello FM Dennis Monokroussos
Everybody knows the most important aspect of chess is the ability to perform calculations, thus the reason why we all solve puzzles. However, nobody ever says which type of puzzles is better. I normally solve 100 puzzles, from [website], correctly each day to improve, but most of them don't feel like real life situations and usually have some sort of clue. Should I continue doing these puzzles that are fairly easy, or try going through the carefully selected puzzles from [tactics book]?
Thank you for reading my message, posting so many instructive videos, and for updating the blog so frequently and with such interesting material.
Perhaps it isn't addressed as often as it should be, but it's untrue that "nobody" ever discusses which kinds of puzzles to work on. For one thing, I've addressed this topic before, and I'm sure I'm not the only one to do so. In fact, I've addressed the issue fairly often, and probably will again. The following is a brief summary of some of my thoughts on the matter.
Basic tactical competence (definition): When I say that a player has basic tactical competence, I mean at least two things. First, they've reached a point where it's very rare for them to blunder mate or hang pieces. Maybe it happens sometimes, especially in time trouble or in a very difficult position, but it's not a besetting problem. Second, the player is familiar with basic ideas like generic double attacks, knight forks, pins, skewers, smothered mate and so on, and is capable of finding and applying such tactical themes in their games.
Step one (for beginners and tactically weak players): For players who lack basic tactical competence, the biggest need is to improve one's board vision and to learn elementary tactics. Simple, rote learning is very useful here.
Step two (for average club players who have reached basic tactical competence): Overlearning elementary tactics by continuing to practice them on a regular basis is useful, but it's time to move up to more challenging positions. (These can be positions that don't fit into the obvious categories, or if they do, the way in which they do isn't obvious to someone who has been working with the basic books.) Thus you consolidate your gains, the ideas you already know, but start to stake out new ground. This should start happening around 1500-1600 USCF/FIDE. (I don't count internet ratings here, which are very hard to calibrate with tournament ratings.)
Step three (for strong club players): The occasional look at elementary tactics is still worthwhile, to keep sharp, but the focus ought to be on more and more challenging material. If step one is 100% basic tactics and step two about 50-50, maybe give or take 10% either way, I think step three should be 80+% challenging tactics. Even this should be somewhat fine-grained. It's not necessary that all one's tactical work be migraine-inducing; I'd say around 1/4 to 1/2 of the hard work should be extremely hard. As with actual playing, a challenge is good, but if there's almost no chance of rising to the challenge it's good to aim a little lower most of the time.
As for particular works, I've listed them even more frequently, and don't really feel like doing so again at this point - interested parties can search my blog for specific resources. I'm also not interested in repeating myself this time around because it doesn't matter that much: there are tons of great tactics books out there. The key is to use them!