The interesting:
* The white dotted divider lines on freeway lanes are 10-15 feet in length. They look shorter because of an optical illusion caused by the speed of the car.
* We judge speed by the rate at which objects appear to increase in size and the distance at which we can make judgements about speed becomes shorter and shorter as speeds increase.
* The more dangerous a road *feels* the more safe it is in almost every instance.
* "risk homeostasis" - after making any aspect of driving safer, drivers tend to respond by driving in a more dangerous fashion. After a safety device is implemented by car manufacturers or road builders accidents dip down and then slowly rise back to their previous levels.
* making eye contact with cross traffic is more likely to get you hit by a car as a pedestrian.
* purposely driving at slow speeds in stop/go traffic will eliminate the stop/go effect for the entire lane.
* most congestion is caused by people following at the wrong distance. Having two cars per ten with automatic cruise control systems (which adjust speed to account for the vehicle in front) would eliminate almost all traffic jams.
* Adding additional highways and more lanes for traffic tends to increase traffic and erode most of the gains that the extra space would provide.

The awful:
* Almost a third of the book is spent bringing up evidence and research that supports a conclusion and then introducing confounding factors to problematize that conclusion. There is never any discussion of how significant a problematizing factor might be and no discussion as to the range of effect that any un-accounted for factor might affect a conclusion. The result is that the book has the feeling of listening to an old man hem and haw for 200 pages without resolution.

* Lots of generic opening lines that openly make assumptions about the reader: "Many of us have probably...", "we all...", "It's a common occurrence...", "you're probably wondering...", etc

*The end notes shouldn't take the place of the bibliography.