Fair enough, I simply disagree. To me, it is more fun to make a few checks when crafting a magic item to determine how much time it takes and how much it costs, within a reasonable scope, than to just say, "it takes 6 weeks and 800g". In my game, that sometimes means rolling a check with results based on degree of success, and other times is simplified into a single die like the heat mechanic. How long does it take to hear back from a research contact? Roll the die. If you are doing something to speed it up, you might roll a different die or something, but even if the result doesn't actually matter, IME it's simply more fun to slightly randomize the result in a way that involves physical engagement with game materials.
So, this is where we get to my perspective on 5e (other games fall into the same basket, I'm not really picking on 5e). What is the point of these checks? Someone makes up some arbitrary numbers, "6 weeks" and "800g", there's very little to support that they have any special meaning, that 6 is better than 5 weeks in some way. The way in 5e the GM just sort of says "OK, make a check now." but we don't really know what it means to succeed or fail especially, or how that fits into the greater picture of achieving some goal. Its basically arbitrary. Then the inevitable question is does it even matter?
So, with your example of the magic item, WHY is the character making it? Will they fail in their goal if they can't finish before a certain time? That would be interesting to resolve, and the fair way would be some sort of mechanics. So, to me, that's part of some sort of challenge. But if 6 weeks is just 'color', it doesn't change anything much vs 5 weeks or 10 weeks, etc. then I'm not really seeing the value of using dice, or whatever. Again, feel free at my table to toss some dice around, and you can call it a 'crafting check' if you are so disposed. Presumably the game is going to proceed at the point where you're done, one way or the other.
I'm probably in between, on the issue of mechanically delineating stuff that could just use the same mechanic. Some stuff feels completely redundant to me if it isn't somehow mechanically distinct from other stuff.
For instance, I don't think that any game that wants to do both physical challenges and social challenges well, should do them with the same exact mechanics. They should feel different in play, not just in description. OTOH, I don't see any real value to getting nitty gritty with a hundred different weapons that each have their own mechanics.
I would think these would definitely engage different traits of the PCs and need different resources, strategies, etc. Actually, in my current rules a 'Physical Challenge' (IE a discrete situation of physical danger which is resolved in the moment) is an 'Action Sequence' and uses the combat rules. It MIGHT technically not involve actual fighting, but the same 'move around on a map and make checks to take specific actions in an initiative order' can still be used. A social challenge OTOH is just plain straight up an SC. When actual combat is involved the distinction is entirely clear, but I admit that other sorts of dangerous situations are often better handled as SCs too, or even just "a check during an SC" depending on scope. This is one area where I'm quite flexible, I generally break down whatever the plot is into a challenge that has a few checks in it, but clearly the more central something is to the player's primary interest/focus in the game, the more detailed you would normally go.
So, if there's some side quest to help the clerics deadbeat cousin get out of hock with the local loansharks, that's probably just a challenge, the equivalent of an encounter, even if it involves going to several locations, intimidating some guys, and paying off a debt, etc. The players are interested enough to do it, but the campaign is mostly about demons, or whatever, not loansharks. That's kinda how I roll nowadays. For that kind of game, its much better to have a ruleset which deals easily with actions at all scales.