They didn't claim that. They claimed that any (sound and consistent) finitely axiomatizable theory (basically, any recursively enumerable set of theorems) can only prove finitely many theorems of the form BB(n) = N.