The much anticipated plans for the Space Launch System (SLS) were unveiled today. It will cost (at least) $18 billion to get to the first unmanned test launch in 2017 of a 70 ton payload version and, according to Space News, the launch rate will be one every two years. If the system lasts for 20 years, that means a total of 10 launches. If the system cost nothing after the first test launch, the SLS would cost $1.8 billion per launch -- more than the space shuttle.
Now suppose we spent that money on private launch services. Right now SpaceX is advertising $125 million per launch for the Falcon Heavy, which will lift 53 tons. That means we could buy 144 Falcon Heavy launches with the money needed for a single test launch of the SLS.
Furthermore, in a letter to the Space News editor on February 7 SpaceX president Gwynne Shotwell said: "we .. develop a heavy-lift launch vehicle with a 150 metric ton to orbit capability ... We can do so for no more than $2.5 billion, within five years, on a firm, fixed price basis with payment made only on achieving hardware milestones." The SLS is eventually supposed to lift 130 tons, so for 1/7th of the money and a year's less time SpaceX thinks they could do the job.
But perhaps SpaceX can't! What is the record? SpaceX has developed two vehicles in the last few years, Falcon I and Falcon 9, both of which have flown successfully. In the last 20 years NASA has tried to develop four vehicles, the aerospace plane, X33, Ares I and Ares V. None have made it into space.
Of course, the SLS is intended to be human rated, however, the Falcon series is designed to be human rated too.
The short story? From the point of view of space development, the SLS is fiscally insane and there is no reason to believe it is technically superior.