The price tag is not a bug, it's a feature. What, you think DARPA would pay you millions of dollars to invent a real mule? Buck Rodgers movies didn't have mules!
I will say, though, that compared to the typical military-industrial spendthrift extravaganza, this project looks really worthwhile. Think: powered prosthetics for quadriplegics. Think: remote exploration of Martian mountains. Think: lots of generally applicable algorithms for real-time physics.
Because this is only the beginning, and soon you'll have versions that carry ten times more than a mule, travel faster, and can be controlled remotely.
Interesting question. Perhaps we will witness a renewed proliferation of domesticated animals in this millennium, a sort of "domestication 2.0". Robotics may actually help inspire it, along with our knowledge of so many more kinds of animals. As robot designers study lots of different critters to improve their products, people will start thinking, "Hm, some of these animals have skills." The range of animals we've made use of so far is probably tiny compared to the range that could be useful. And our ability to breed or engineer them in desirable directions has improved a lot. So yes, maybe there will be real mules too, or some other kind animal we haven't thought of yet. (Have you ever seen guide horses? http://www.guidehorse.org )
I will say, though, that compared to the typical military-industrial spendthrift extravaganza, this project looks really worthwhile. Think: powered prosthetics for quadriplegics. Think: remote exploration of Martian mountains. Think: lots of generally applicable algorithms for real-time physics.