She falls into the same trap that most do who think about this problem -- that the super-universe in which our simulation is embedded has physics anything like what is being simulated for us (or with us as a side effect.)
If it can do uncounted googleplex operations per second, with similar abailability of storage space (or equivalent for an analog computer!) then none of her speed concerns are valid. Indeed, the cosmic speed limit here is a curious oddity, perhaps deliberately ala Vinge's Zones of Thought.
As for Bell's inequality and hidden variables, again, if it is all simulated, none of that matters. Hidden variables is only an issue if you need to maintain Einstein's concept of reality, that there are real objects "out there" with real, measurable properties. If one gives up on that reality, one can base quantum mechanics on a deeper classical realism with no problems whatsoever.
But even that need not be the ultimate reality. But her concerns are only issues if one, needlessly, and I submit oddly, wants to maintain that that parent computer's physics is anything like the physics being sinulated.