I completely agree, Piety should be weaker if taken as starting tree, my gripe is with Tradition. I actually don't mind Piety opening in Ancient era, I regularly open Piety after the first couple of policies in Tradition/Liberty to get the production bonus to Shrines at the time I actually start building Shrines. Rather, one could argue that the AI should be programmed to not max out Piety straight away.
But then again, that's one of the dilemmas of which role the AI should play: Should it always play optimally, and should it sometimes play sub-optimally in order to challenge the player in a specific area? For instance, the fact that 1-2 AI players almost always max out Piety directly puts some pressure on the player if you want to target specific beliefs - for instance, getting Jesuit Education is rather rare because the AI will almost always get to it before you while you're busy filling out Tradition/Liberty, and I think that's fine. So basically those AIs sacrifice themselves to spice up gameplay for the human player, and I guess that's fine (as long as some of the other AI players play optimally).
Mmmm the other annoying seeing a civ get Jesuit Education and Cathedrals at 2000 BC. Hence the other reason why I favor Piety being delayed.
Usually I tend to ignore early Piety - if I get it it's usually to get Temples more quickly. Shrines can manage by themselves and on higher difficulties you generally need more than Shrines for a religion anyway