Platooning has been talked about for years now, but there are a number of barriers to its success that simply can't be answered.
1. The energy savings just aren't great, unless you have the cars stay right behind one another.
2. Lane changes would be extremely dangerous.
3. Only the single rearmost car in the platoon can safely leave; others would need to be booted out to the side in order to leave the platoon. This is a huge limit on capacity. The article says you would open a bigger gap, but the problem is that you can't switch the car to manual in the middle of the platoon, but you also can't wait until the car is out of the platoon to do so since then it would be mingling with non-automatic cars and is at risk for a collision.
4. Pricing and a workable business plan are lacking; having it as a free service would be prohibitively expensive (but highly ironic in America - "trains are a waste of government money, but ROAD trains? Now that's a different story..."), and a pricing plan for the service would be difficult to implement.
5. Platooning works best on relatively empty roads, of which there are a decreasing number nowadays. On congested roads, wireless interference (both in terms of noise, and the potential for hostile interference, such as hacking the wifi and having the cars crash into a wall or something.
6. You're reliant on the competence of the driver in front; this is the case for all transportation modes. The problem is that on roads, professional drivers are forced to intermingle with unprofessional drivers, which means it's inherently more dangerous. If the lead car gets into a crash, every car behind it would also end up in a chain reaction collision (given the number of cars, all such crashes would have high fatality rates).
7. Auto-drive is problematic since different types of cars have different characteristics. While smooth driving might be easy enough, sudden braking or swerving to avoid hazards would mean the computer on the lead car would have to suddenly calculate a long list of variables (road conditions, speed, handling characteristics of cars in platoon, etc.) in a very short time, and for a number of cars.
8. Compatibility would hinge on dedicated software, massive overhauls for existing cars (or purchasing a new car), and new internal hardware. The cost of the software for a wirelessly controlled autopilot that also is highly reliable and precise would be in excess of $5000.
9. If one car's auto-drive broke, the whole train will likely break; even a decent safety record would require multiple redundant backups and a fail-safe device, which again raises the cost.
10. This might work on a smooth, dry, and somewhat empty test tracks with limited radio interference, but conflicting traffic, radio interference, poor roads, and weather issues all make this much more difficult.
So in short, no.
I'm Batman!