I don't hear much about GWT any more.
When it first came out, it seemed to be all the rave in certain RIA circles.
But lately my impression is that GWT has suffered a dip in popularity because solutions like Jquery/MooTools/Prototype offer a much easier way of solving the same problem.
Is this accurate or does GWT solve a different type of problem?
Well, sort of.
The end-product of GWT ends up being something similar to what the other JS frameworks provide - a JavaScript solution that, for the most part, abstracts-away all the browser and DOM inconsistencies so you can focus on Real Work™.
However, there is a significant difference between GWT and the JS frameworks - programmers who author code with GWT are writing Java, which is then compiled into JavaScript.
So what it really comes down to is this: because great JS frameworks now exist, how many web developers are left that know Java, but don't know JavaScript? Ergo, don't know jQuery or MooTools or whatever?
To me, the answer to that question is: not many.
GWT and javascript libraries serve two different purposes. GWT generates web applications from java code and javascript libraries can be used as a component within web applications. Because javascript libraries can be used in many different types of web application projects (ASP.NET, Ruby on Rails etc.) they may have more of an audience than GWT but the functionality of GWT isn't something javascript libraries completely replace on their own. Depending on needs, both can be very useful for their respective audiences.
GWT of course is for folks who like to code using Java. If you are not a Java geek, you will not have passion for it, nor will be able to appreciate it. It has completely eliminated the needs to learn one new framework. All you need to know is Java, that's it. js frameworks as someone already pointed out can be used across platforms (.NET, Java, Ruby), but I am not sure if this statement is actually true.