Like many ideas this proposed idea from Amazon is simply that, an idea. Look through the backlog of ideas or inventions that have been patented and you see a lot of 'great ideas' that frankly are ludicrous if you look at them now...and what they propose is not going to happen in the short term, I can guarantee that.
However, it is doable, but requires some fundamental things that don't exist and may not be able to be overcome:
1)Given Amazon's model, that rail would even work for them. With the 1 day delivery or intra day delivery they are banking on having the inventory local, they have local fulfillment centers that allow rapid delivery using the current model (though I have had 1 day delivery on items that were shipped from across the country). The problem with rail, especially in this day and age of "precision railroading" or whatever they call it, is that rail shipping if I read right in trains magazine, has gotten even less time conscious.
Does Amazon think that they are going to have a train full of shipping containers full of various items (sort of like their current distribution centers) and as people order, they check where the nearest of these trains is, and ship from there (the description sounds like that)? If so, then Amazon's plan is kind of like the 1950s idea of having trains with missiles on them continually running the rails, in their case it will be having trains with Amazon containers running a kind of continuous route and having trains all over the country doing this running some sort of regional loop(kind of like running around one of our layouts).....and I wonder if they plan to restock the containers on the fly, too...
Just don't know if rail could support this, if traffic would allow it or if the railroads could handle having trains running like that.
2.The future of drones is cloudy as well, given how much Amazon delivers on a given day, just not sure having so many drones working in local airspace would work. Too many chances of accidents, Drones being blown into power lines, Drones dropping packages from a height and killing someone because of a malfunction, and to be honest, swarms of drones in neighborhood delivering packages sounds very intrusive. Sure, you get deliveries in neighborhoods all the time via UPS, Fedex, the USPS, Amazon, but it isn't the same thing of having airspace full of drones, very different level of annoyance and yes, risk.
3)Can this be run by AI? The self repair by robots is still questionable, there has been limited success with robotic and self repair, but it isn't quite science fiction, it is just really early. Likewise, a distribution system like this will require sophisticated tracking and dispatching capabilities to make it work, and as good as AI/machine learning systems can be in this area, it isn't there yet....but could be, and not that far in the future.
AI systems don't need to have emotions to function, they don't need wit or humor to be able to solve problems, they need sophisticated self learning systems to learn as they go the way people do, things like intuition and human's ability to put together a bunch of guesses and inference to solve a problem are not easily replicated by AI, but they don't have to be, AI intelligence is much like work on animal intelligence, what they have found is that animal intelligence uses very different areas of their brain working in different ways to do things humans do very differently; AI wouldn't replicate human processes, it would find a different way to get the same results, and has.
Put it this way, AI is used all over the place in ways people don't even know, it is used in everything from appliances and a lot of the trading going on in financial markets have AI components to them, including machine learning, it is used to solve problems that humans haven't been able to because it works differently *shrug*. Rule based AI has been around a long time, what is being developed right now is not rule based, it is based in self learning and in mimicking how humans solve problems, and it has advanced a lot in the last 5 years or so from what I can tell.
Does this mean Amazon's idea will work? If it does, it has some fundamental problems that would definitely block it today, and not all of it is technical in nature, there is a lot of human engineering that would allow this to work, including the railroads, not exactly paradigms these days of innovative thinking or thinking outside the box, and in people accepting this kind of system. I agree that Amazon's current model likely is not going to work long term with shipping, it works now because Amazon makes money a lot more from its cloud businesses than the retail side, but as others have side, they are expecting it, like their warehouses, to become more and more automated, to really make money with their model of almost on demand shipping.
