I think the short answer to that question is: it depends on how you play.
The jump from 2nd edition to 3e was definitely about revitalizing the game by discarding accumulated cruft and fixing things that were broken. No more THAC0, yay! And then 3.5 was an incremental improvement to 3e that fixed a bunch of bugs and tweaked things to make them run more smoothly.
Whereas 4e is not so much about fixing the previous version as it is shifting and tightening the focus. I'd say it backs off from doing a huge variety of things adequately, in a mostly simulationist, way and instead does a few things quite well, in a somewhat more narrativist way.
4e works a lot better with structured adventures than it does with free-form improvisation. The skill challenge system still may have some mathematical issues (I'll post my homebrew version once I've tested it a couple more times), but the idea behind it is really solid. It doesn't really support adventuring with a party of five 0th-level cooking students, but it makes sure that in a standard party, everybody's got something fun to do most of the time. You may not be able to fake up a tengu ninja as easily, but boring old human wizards are now a functional option -- even at 1st level. Et cetera.
For some people, that adds up to a huge improvement. For others, it's a big lose. A lot of people will be somewhere in the middle until the find their footing and figure out whether they're comfortable with a play style that's suited to the game.
I what a prospective GM needs to do is to read through the books with an eye toward understanding what the design is aiming at. Run an adventure that actually follows all the advice given, to see how the system drives. And then think about whether that's a good fit to the kinds of game you want to run and your players want to play.