I prefer having the smarts in the charger too.
It's not clear if you mean instead of, or as well as, smarts in the battery.
If it's the former then that would, IMO, be a bad move since the charger would need to know about the battery. Use the wrong charger and it could be goodbye battery (and whatever is nearby).
Plus, it makes chargers more versatile for different devices.
Actually, less versatile. Consider that the charger would need to know about everything it might ever be charging, which is an impossibility. So there would be a subset of everything, and the likelihood is that the subset would be quite small. What happens about the stuff the charger doesn't know? Not awfully versatile with those, I guess.
On the other hand, a 12VDC supply is pretty versatile and generic. Doesn't give a damn about what's sucking juice from it, won't try to control whatever it is erroneously, etc. Of course, that means the batteries would need smarts, but once all of them have it the cost of adding it is very small. As an example, once upon a time a simple one-axis electronic gyro cost serious money, eventually falling to "er, I'll think about it" cost. Nowadays you get 6DOF sensors for peanuts, mainly because every phone contained a gyro or accelerometer or, more often, both.
I think things like charging, where there are two distinct entities that share nothing except the connection, should conform to Postel's law: "an implementation should be conservative in its sending behavior, and liberal in its receiving behavior". Perhaps USB PD does that.
But it seems like the industry is moving towards USB-C for uniformity.
I think uniformity in this context is a Good Thing. The problem is that USB-C is quite complex, needing even special cables with smarts in them to really do the business. But it falls back to something sensible if you're cheaping out and just want simple.