The number of abstractions/libraries seems irrelevant to the discussion if they can indeed be out of your way (you're saying because of all the boilerplate etc. that they aren't, so I guess we're talking about different things). If they're powerful and there's a small surface error for dealing with them then great, that's a high return on investment situation.
Getting to that small surface error takes an IDE sometimes which is unfortunate on some level but I don't find anything fundamentally wrong with it.
Personally, I can't stand even simple programs in Java with a text editor because of the amount of subdirectories you need to suddenly be aware of... org/blah/blah/blah/blah/blah (this is how I learned it originally).
Maybe things have bettered since then but from what I see at least the maven disease seems to still be all over the place. I'm not saying that all these things are necessarily bad. I just grew really tired of juggling monstrous xml blobs around while still being forced to write (or generate) metric tons of boilerplate code.
Same here and imho that's quite an alarming indicator for the "towering abstractions" that paul mentioned. When you can't realistically navigate a significant codebase without tool support anymore then many of the components just aren't "out of the way" as much as they used to be or should be. Personally I jumped ship when java went through the heights of its XML disease which is apparently cooling down nowadays. The available toolchain (part of which you mentioned) is mighty impressive, no question. But it's the amount of boilerplate- and meta-code required to glue the stuff together that's driving many of us away.