I know the context of the article is using these signs in urban areas with closely-spaced, timed traffic signals, but in reality the application for these signs could be much broader than that. It's a nice idea to let drivers know how slow to drive to never have to stop, but you're still going to have some people jockeying for position and racing from red to red, and that will tend to screw up the whole dynamic of the traffic flow.

It certainly makes a lot of sense to use dynamic speed signs since driving conditions are dynamic. The signs could perhaps be controlled by computers or DOT staff that monitor air/road surface temperatures, weather conditions, traffic speed/density, visibility, and other relevant factors and reduce speeds for traffic approaching slower or more dangerous conditions well ahead of time (especially useful for rural stretches of highway). While less flexible, these signs are much less expensive than installing overhead electronic reader boards on highways (plus they're made by a Portland-area company).

This is a specific example of the general family of management techniques know as "Intelligent Transportation Systems" and collectively they are a way to wring more capacity out of the current road system without adding lanes.

The problem with this kind of light progression is it doesn't work in heavy traffic.

On NE Broadway the lights are set so that you can drift down the street at at something like 25 or 30 and hit every one. However, if you do so, the car in the next lane speeds up and gets in front of you at the next light. In addition, cars are always entering the street and waiting at the lights. If you try to coordinate your speed with the lights you end up missing one every few blocks.

On Grand Avenue, all the lights in a multi-block stretch go green at once. That produces only slight speeding, and it flushes the traffic right through.

Different street require different light progressions.