Been a while since I read this paper, but:
The main criticism in this paper is, that the problems are asked and preprocessed in a way that makes them almost trivial. Basically, CPU manufacturers could boast about 10000% faster prime number factorization, if all they were doing was prime number factorization of 2^n^
They even suggest, that instead of solving known and borderline trivial problems, quantum computers should instead be evaluated using random problems.
^Also, they should thank their dog in the acknowledgement section^


I remember having recurring dreams in which I would fall into the gap in the elevator or between the subways/trains and the plattform, etc. and then jolt away with a whole-body flinch.
The dreams went away as I got older, but I still grip stuff (e.g. my phone) as tightly as I can if I have to carry it across those gaps ^^