My thoughts exactly, I didn't see anything on the paper addressing that.

I suppose it's assumed that with the contributions of this one, future work can be done to make it faster.