Why isn't mathematics too big for the human mind? Wiles only baaaarely got Fermat's Last Theorem. Nobody has Goldbach's Conjecture yet. And these are just the things we care about! What about the arbitrarily large number of things we don't care about?

Why is Mary, with her human limitations in transference between declarative knowledge, procedural knowledge, direct perception, etc. etc., treated as the gold standard for what is possible with a computational device? Why is our imagination of what it would be like to be Mary considered a useful approximation to Mary herself?

It's really easy to win an argument when one gets to make a bunch of unjustified assumptions. I know the assumptions aren't original to you--but this doesn't make them any better justified.