Do you really consider it a train wreck? There seem to be some useful new features.
The tragedy, in my opinion, is that we don't have more to show from a decade's work by so many brilliant people. C++ is straitjacketed by ISO standardization.
Design by committee is a tricky business, especially if Bijarne says they are always aiming for the consensus and not a simple majority vote.
Even not considering C++, this approach is more likely to result in the acceptance of a feature that everyone feels equally "meh" about rather than something that "only" 80% really likes, but 20% opposes.
It was a train wreck beforehand. It has since gotten even worse. The remarks I made at http://news.ycombinator.com/item?id=598805 still stand, IMHO; basically due to backwards compatibility concerns, there are almost always at least three ways of doing any trivial thing in C++, and the interactions between them are terrifying.
That's a feature, not a bug. Code you write in C++ today will probably compile and run 30 to 50 years from now with very little (if any) modification. Same is true for C. These are serious languages used on major projects (Joint Strike Fighter, Mars Rover, etc.) by governments, militaries and space agencies. They do not evolve quickly and when they do it is only after much careful deliberation. Don't get me wrong... I enjoy the new, fun, expressive constantly changing languages as much as anyone, but when I'm doing serious, mission-critical code, I use C++.
C++ has many compilers for loads of embedded platforms. It has to be standardized. It's not comparable to the languages with two or three implementations that all sit on top of Windows or UNIX. It is much harder to change C++ without hurting many compiler writers.
The tragedy, in my opinion, is that we don't have more to show from a decade's work by so many brilliant people. C++ is straitjacketed by ISO standardization.