What the article doesn't state clearly is that this assumes the strong exponential time hypothesis: it assumes that SAT cannot be solved in time 1.9999^n -- in other words that it's impossible to do better than the brute-force algorithm, which has complexity 2^n (up to polynomial factors). That's a very strong assumption.