But yep, I agree - the author shouldn't necessarily try to start a debate without offering specific use cases and failures. Far too much conjecture IMO.