I don't think the author was claiming that significant number of people hold those assumptions. I thought the assumption was that the most people implementing QF believe that QF is optimal for their use case. However, the author's observed use cases tend to not match the preconditions for QF's optimality.