My logic goes like this: First, we have planes today that can take off, fly and land without a pilot. Pilots are there only in case something untoward happens. Second, we already have remotely controlled aircraft, which we've seen, for example, in the Afghanistan war. Third, computers themselves, over the next 18 years -- if we stay on this Moore's law kick -- will be about 4,000 times more powerful than they are today.
The next step in this logic is that with computers increasingly a part of critical infrastructure, the industry is going to have to focus a lot more on making machines that just don't fail. If we go at that hot and heavy for five or ten years, I imagine arriving at a methodology for system design that yields as much dependability, on an everyday basis, as the triple-redundant computer that flew guys to the moon. Finally air traffic control will no longer be based on staticky communications with people staring at radar screens; it'll be completely computerized and recast.
The assertion "Commercial air carriage will be pilotless in the U.S. by 2030" can't be really true.
No licensed air carrier (commercial or private) will be able to use it without at least one pilot supervising the whole process in the pilot seat, even though the technology to take off, cruise and land automatically already exists.
On takeoff, the training and timing around handling emergencies such as engine failure at rotation are not going to be transferrable to autopilots and machines.
On landing, the automated airplanes have to sequence in with many older airplanes with human pilots. The towers and air traffic controllers love to change everything at the last minute, and adding the ability to make changes by computer while simultaneously using voice is not realistic. Finally, the FAA changes so slowly that if this were even all possible, the adoption and certification would all take at least 50 years.