[TheLoneWolfling]

It's not an ISP of infinity.

You're forgetting mass-energy equivalence. <x> energy / <y> thrust is a specific impulse, once you work in the factor of c^2 (that being the minimum amount of mass you need to burn to get the energy to produce the thrust.)

In this case, for instance, they are saying they think it produced 20 uN with an input of 700w. That works out to an effective specific impulse of 2.5x10^9m/s as an upper limit (yes, this is tachyonic. One of the problems with this drive is that it throws conservation of energy and momentum out the window...), assuming we could convert mass to energy directly. (If you're using D-T fusion, it's a specific impulse of ~9.7x10^6m/s instead as an upper limit, assuming 339.72 TJ/kg. Other reactions are lower.)

If you're assuming solar panels or some other external power source, then the only real difference between this and a laser is that this produces a constant factor more thrust per energy input. Sure, it's reactionless - but, effectively, so is a neutrino source, for instance. All that means is that you don't need to be as careful as to which direction you burn in.