I don't like innovation in filesystems. ext4 might be boring, but ext4 wrecks don't make the evening news the way wrecks with ZFS do. I worked on a system that used reiserfs in production and we were always dealing with problems caused by the weak reliability guarantees. For instance when the system crashed we'd find our filesystem was full of files full of trash data (it allocated space for the file but never wrote the data so this is what was left behind.)
Since the system would try to read the trash data, this was a problem.
I was looking at the btrfs documentation the other day and noticed that (1) it behaves worse when space runs out than ext4 does, and (2) there's no accurate way to measure free space on a btrfs volume so it's hard to avoid running out of space.
Whatever benefits btrfs has is erased if you have to deal with wrecked and full filesystems all the time.