In the blog L.A. traffic sucks: Let's fix it! the author argues that driving is an "addiction". I believe people are behaving individually rationally (with perhaps irrational preferences) to achieve personal satisfaction, which may result in societally unoptimal outcomes.
Italianesco, the author, asks "In other words, if I understand well, the speed of traffic is directly proportional to the cost of gas? The higher the price, the higher the speed (fewer cars on the road); the lower the price, the lower the speed (more cars on the road)? And the faster the speed, the sooner you get "there" and the more time you save? " I think Italianesco understands correctly. The empirical question is the degree of relation between those variables, not the direction of the effect.
He also takes a potshot "Academics love calculations right down to the last digit." What is the alternative to accurate calculations? Inaccurate calculations or simple assertion. We need data to understand people's response to higher prices, data like the National Household Travel Survey or analyses like those referenced in this paper.
"Academics and staticians interested in numbers might want to find out what are all those people doing on the road at ALL TIMES of day and night in L.A. One of the best solutions to the traffic problem may lie in the answers to that question." Yes, that is of course non-work travel, which is no surprise to transportation academics (and one hopes, not to professionals either). Most travel is not the commute from work to home.