*Well, the term "mistake" offers the presumption that you could have successfully done something else... Like, was the Hitler Vengeance-Rocket a "mistake"? Was it a "mistake" to launch Gagarin into orbit? Every regime that ever threw money and credibility at space rockets has come to a bad end, and the more committed they got, the worse off they were.
*The Shuttle likely doesn't compare with the severe "mistake" of the nuclear arms race or the internal combustion engine. But yeah... the Shuttle was a massively expensive technical boondoggle, and it always was. Now it's one with the Superconducting Super Collider, and it's gonna be interesting to see mankind trying to mentally deal with the extinct Shuttle as a Gothic High-Tech icon of atemporal retro-futurism. A dead cow never reeks quite as much as a dead sacred cow.
http://www.technologyreview.com/computing/37981/
"Forty years ago, I wrote an article for Technology Review titled "Shall We Build the Space Shuttle?" Now, with the 135th and final flight of the shuttle at hand, and the benefit of hindsight, it seems appropriate to ask a slightly different question—"Should We Have Built the Space Shuttle?"
"After the very expensive Apollo effort, a low-cost space transportation system for both humans and cargo was seen as key to the future of the U.S. space program in the 1980s and beyond. So developing some form of new space launch system made sense as the major NASA effort for the 1970s, presuming the United States was committed to continuing space leadership. But it was probably a mistake to develop this particular space shuttle design, and then to build the future U.S. space program around it.
"The selection in 1972 of an ambitious and technologically challenging shuttle design resulted in the most complex machine ever built. Rather than lowering the costs of access to space and making it routine, the space shuttle turned out to be an experimental vehicle with multiple inherent risks, requiring extreme care and high costs to operate safely. Other, simpler designs were considered in 1971 in the run-up to President Nixon's final decision; in retrospect, taking a more evolutionary approach by developing one of them instead would probably have been a better choice...."
