SystemD was buggy at the beginning and had some strange defaults, such as binary logging, which have been fixed. It's unfortunate it has a spiky developer and that it was released the way it was, but I'd say 110% of the issues people complain about have been fixed, while sysvinit remains the init system of choice for those who are happy to accept they'll have to use rescue disks occasionally if they have an entirely reasonable configuration that just doesn't work because of timing issues or network resources being unavailable.
PulseAudio... I was never convinced that putting all the audio stuff in user space was a great idea, PulseAudio exists largely because the kernel developers didn't have the right skillset to understand what people needed from audio, so someone stepped up to write something as a stop gap until they figured it out. Again, like SystemD, unfortunately early versions were bug ridden and it had a spiky developer - the same one! - behind it. But it wasn't per-se a bad idea, and ultimately it ended up working, it just was a hack to deal with a greater underlying issue that nobody seemed to want to fix.
Wayland - just a shit idea. Made by a group of people who joined the X11 project long after the original people had left, whined and bitched about having to maintain someone else's code, and then decided instead of rewriting it to create some utopian replacement. That it even ended up with something after 15 years is a fucking miracle because generally when people say "Oh this working, complete, system is imperfect, let's replace it with something perfect" they all have different ideas as to what is imperfect about it. But, no surprise, Wayland is garbage. It throws out 40 years of proven design to create a system that's only "modular" because everyone scoped parts to things they wanted to work on, rather than because they were a good idea. Wayland advocates frequently make claims that are verifiably false: "Screenshotting works!", "We created network transparency" (no link but you notice that claim died pretty quickly - no support in ssh like there is -X for X11, wonder why...), "X11 is inefficient, Wayland is the best for games", etc, etc.
Wayland is being pushed because the devs, who control X11, IBM, who funds everything, and so on, cannot be seen as wasting the last 15 years on a side project that didn't work out, leaving X11 to fester when virtually everything useful in Wayland could have been put into X11. In mean time, the only reason it's popular is because people have been repeating the same smears against X11 since the late 1990s about it being "inefficient" and "slow", which sounds credible when you find out it has network transparency and communicates over a socket... until you discover that, no, the only time you see large blocks of data sent over the socket is when it's actually running over a network.
This truly sucks. And no, I think XLibre's leadership is going to drive away many of the people who could fix this. Things just keep getting shittier.