You're right, I was only looking at line (5) briefly and thought he was talking about the minimum change of entropy. S ≥ k ln2 is incorrect, as you say. A syst

You're right, I was only looking at line (5) briefly and thought he was talking about the minimum change of entropy. S ≥ k ln2 is incorrect, as you say. A system whose state is fully known has no entropy. One can either see this through counting the number of microstates (which is 1) or by applying Shannon's definition of entropy to the trivial probability distribution which is 1 for precisely one state.