I think one way to falsify simulation hypothesis, assuming universe running simulation have similar physics, is by finding limits of computation required to run the simulation, at some point we can find a bug or a glitch