I think because it's not great against tokens and not great against Outcome. And then if your hand is only Spheres and wire, no threats, that isn't enough. Shops has to kill fast. You'd have to not cast a threat in order to cast a Wire. Also since games are so quick the shop deck can be bottlenecked on mana, so you'd be tapping your own creatures which is counterproductive. Then it fades away and your -1. It would be like playing Tangle Wire in white weenie
In the counter argument:
I ran three mainboard in a Vintage daily a couple of weekends ago to a top 4 finish. It was frequently the card I was happiest to see with the exception of a starting hand with two of them which lacked other relevant action so I had to ship it.
It was wonderful in the mirror match.
It wasn't terrible against mentor.
Wire is one of the casualties of restricted Lodestone Golem. In a world where you can reliably cast a turn 2 Golem and follow it up with a Wire (un-taxed, mind you) it was basically a game ending play, especially game 1. Now that Shops needs multiple creatures to obtain the same clock, Wire is less appealing.
It seems like current shop decks are trying to push their fundamental turn ahead of the mentor decks instead of delaying their opponent. Before the Lodestone restriction, I experienced an amazing matchup against Mentor with the Tiny Robots deck partially because of this: who cares about a mentor played on turn 3 when you have 14 power on the board and an active draw engine? Combine with the fact that the mana denial mirror is now decided by who has the most/biggest dudes instead of who can stabilize their resources and it makes sense why some lists are gravitating away from Tangle Wire.
That being said, I ran one of Jazza's lists recently and I really wanted to topdeck a Tangle Wire all day long.
I wouldn't discount the increased importance of the workshop mirror. A deck that cuts 4 Wires for 4 threats is going to have a big edge against a Workshop deck that doesn't. Even if Wire was better against, say, Mentor and Outcome (and I'm not sure it is), it might still be the least-risky card to cut in order to get an edge in the mirror.
I suspect this has been correct for years, and is just being discovered now, rather than a reaction to some new card pool or metagame change.