Myers writes:

Sure, you can just arbitrarily set the time-scale of the simulation, but then you mess up the inputs from outside the simulation. And you can’t model a human brain in total I/O isolation without it melting down into insanity.

I didn't feel comfortable dismissing his objection out of hand, because I wasn't exactly sure what point he was trying to make. Then I read Carl Shulman's comment, and now I'm thinking it probably just didn't occur to him to simulate the brain in a sped-up virtual environment. Probably he assumed the simulation was expected to interact with the real world as flesh-and-blood humans do, just while thinking faster. If this was the goal, it seems his objection would be valid.