As a practical engine, EM drives are worse than chemical rockets or even ion thrusters (which produce about 25-250 mN/kW, according to Wikipedia).
Obviously the first engine ever exhaustively tested is usually not the highest performance achievable. A factor of a thousand seems possible although a hundred is more likely.
So for a very large nuclear plant, if the EM drive were weightless (LOL) then maybe 1/10th G applied to that nuclear sub is some kind of theoretical limit not entirely ridiculous but possible.
Now consider that to one sig fig a G of acceleration for a year is the speed of light. And if you're careful maybe you could run the plant for three decades.
So a round trip to the speed of light and back again is more or less back of the envelope possible to visit some distant star.
Relativity rears its ugly head and ten years of ship time around the speed of light is a lot further than 10 lightyears and time passing on earth is a lot longer than ten years.
Of course at the speed of light you're going to need a rather substantial hull, an active radar that can avoid interstellar crud, etc. If you can detect "stuff" at a day out, at a tenth of a G accelerating for an entire day you can be quite far away indeed by the time you pass it.
I can't run the numbers in my head like above, but even 1/100th G slices a very substantial amount of time off a Mars mission. Some unverified google results imply 1/100th G means Mars in a month not a year and a half.
At least as a back of the envelope calculation its quite a useful vehicle.
My point is that there are plenty of possibilities to continue to develop propulsion technology.