Most of them aren't actually all that small, it's just that the perspective it's being filmed from (the narrator's head-camera) is on a stage, and VR warps things in your periphery pretty hard.

There's also the fact that it's easier to be immersed if you aren't leaning heavily on realism; Oculus has a lot of research on this, actually. For example, if you have a photorealistic avatar but aren't doing eye-tracking, it's actually more jarring and discomforting than a cartoon avatar without it. That's why the initial avatars they did had force-obscured eyes and no color, and the new ones are literally cartoons: your mind can adapt to something very different, but it can't adapt to something that's just subtly different.

However, some people do choose to do weird stuff. That's mostly a tech thing though. If you go to a normal place, you'll find a lot of actual kids and actual teenagers, and they're relatively normal. Very few people as weird as the tech community.

Really, technical people in digital spaces kind of get a bit wild with their self-expression. Some theorize it's because most of them don't express themselves much in real life.

I've tried this particular app only a handful of times (I don't like using proprietary software, but it's important to keep yourself updated with the state of it), and when I wasn't visiting explicitly tech-related things, it was pretty normal. Lots of copyright infringement of Nintendo characters, as you might expect.