The part that really bugs me, and why they look rendered - is because that is NOT how the scene is perceived when you are actually there. Yes, each separate part of the scene may look like that. But that's not how our brain works. For example on the pier shot: When you are there, looking at the lights - you are not seeing the sky at anything but black. You are not seeing all the detail at your feet. You are not seeing the water as if it were afternoon.
Your iris have contracted so you can see the detail in the light area you are looking at, and that's what you get. Just the same if you were to stare at the ground, water or sky - and let your iris open up to see the detail. You aren't seeing the details around each light bulb, or on the brightly lit ground - it's washed out in your perifrial vision.
Yes, you have a point, but the point cuts both ways. When looking at a scene with your eyes, you actually can only see detail at the fovea, something like 5% of the visual field at the center, outside of that it's shocking how little can actually be perceived, color perception included (it falls of rapidly outside the fovea). Everything at that fovea is in focus, DOF is strictly a peripheral phenomenon.
So in every photo a compromise must be made about where to focus that will not match up with the experience of "being there" and DOF will be made explicit, no longer peripheral.
Unless the effect of only seeing detail at the fovea is pointed out, few people ever notice it because the visual system is not a camera, it constructs an impression of a given scene based on a succession of rapid (and unconscious) saccades. It's an automatic process that continuously scans the environment for salient details. The impression does not include portions without any detail (unless conditions are extreme, like looking toward a bright sun or at a bare lightbulb).
I think it is not a distortion to say that people actually "see" well exposed scenes, complete with detail at nearly every point, deep shadows and bright highlights. We certainly see more in a given scene than a single exposure can replicate.
If you prefer to isolate a single saccade, and compare it to a photograph then surely the photo and the saccade will be roughly similar, but no one actually sees this way. In fact it's extremely difficult to consciously override the automatic scanning system, and stare at only a single point (if you've ever taken a vision test you understand). And even in this case, because we only see detail at the very center, I think you'll agree that a photograph of the same scene looks quite different (in part because you can actually look directly at the bokeh).
So the compromise in HDR is simply different, not better, not worse than a single exposure, but different. It is to move toward duplicating the impression of "being there." Focusing on the result of the way the visual system compiles many separate, unconscious glances into a coherent, well-exposed whole. Certainly it departs from the experience by presenting it all at once (and also by compressing the dynamic range from many thousands of bits to just 8 if it's a jpeg), but I really don't think there's a solid argument that this is any deeper a compromise than the one a sinlge exposure makes. And I maintain that we are simply so used to looking at photos where the details that would be seen are not there due to the limitations of the medium, that a well-done HDR is interpreted as "fake" looking by comparison. This tendency has started to fade for me, although it's much easier with my own photos that with others.
All photos represent a compromise when compared to the scene as seen by the brain. in my mind it's just a matter of which sort of compromise you prefer to make.
EDIT: I want to add that I realize HDR isn't appropriate for every scene, even when the goal is to capture the most detail. Part of the learning process for me is finding when to use it and when it's superfluous. I do find that many artificially lit night scenes can be faithfully captured in HDR.