Although the article didn't speak of how this was done (probably for good reason) I suspect it has to do with using some sort of radio receiver to intercept the signals from both the keyfob and the car. With some knowledge of the algorithms used in the system, it probably wouldn't take much to "spoof" the keyfob.

This sort of vulnerability is why mechanical key with embedded microchop is better -- someone must have both a matching mechanical key, and the digital code. Not only that, but since the communcation between key and computer is done by wire (the key itself) rather than by radio, there's no way to use some sort of sniffer to intercept the communcation.

Also, there's the nonsense of having the car running a receiver all the time "seeking" the keyfob, and a lot more chance of problems such as a car not permitting entry because a battery dies on a keyfob. With a physically linked key, it receives the power it needs for the communication from the car itself, and only when is physically in the ignition. Is it really so much harder to turn a key than to push a button to be worth all that?

A good point that was raised was that both encrption and cracking techniques will improve over time. While in the case of computers, it is trivial to upgrade security software to repair the latest vulnerabilities, upgrading security software on a vehicle ignition will probably be a difficult ordeal -- and another thing to worry about. While top of the line vehicle keyless entry systems could be secure presently, several years down the line, a vulnerability could be found and many current vehicles would still be running non-updated software.