Oops, you're right, I got the math wrong when I looked at the table in my paper. I should have said that scrypt can be over 100,000,000,000 times stronger. ;-)