Right now, IMO, we just need more self-driving cars in general to move the whole concept in terms of acceptability and availability. There is certainly lots of room for innovation, but for now just more product selection alone is valuable.
An example in this vein: imagine a firmware update to a wi-fi router to give it MIMO support. A MIMO antenna isn't any different than a regular antenna; the difference comes in the baseband firmware doing clever-er math to pull out overlapped signals, spacially model their sources, and modulate its own output so the signal will constructively interfere for best performance at the destination.
In a Cliffs-notes sense, both vehicles seem to offer an autonomous driving mode in normal conditions. Like any other competitors, they both have a few unique features that are neat but don't really change the overall experience exponentially.
To be clear, I own neither a Tesla or a Mercedes currently and I'm basing this just on looking over the specs, but they appear to be fundamentally comparable.
In addition the Mercedes' system can't change lines automatically.