You look like an idiot (in a funny way)
VR/AR/MR has been actively developing these years. We have seen Facebook Oculus leading VR, Google ARCore and Apple ARKit leading AR (not particularly using a head-mounted display but I'm sure there will be one), and Microsoft Hololens leading MR. In the end they could all become a similar thing: user wears some sort of headsets (or glasses if powerful enough) and sees a screen with fancy stuff.
If you are the person wearing the device, it is fine. You feel you are awesome. You live in a virtual or semi-virtual world and see things no one else can see.
But, others see you as an idiot. That you are weird (in a funny way). Refer to the following video:
How to make you look less (funny) weird?
The question is: how can you look not so (funny) weird?
While you can share your screen with friends so they know what you are doing, you still look funny, like (didn't mean to pick Lenovo only):


Hmm... see where the problem is? The headset hardware itself!
It is bulky. It has to go around your face. And you look very dull (and thus funny somehow..).
Why can't we make it smaller... just a screen for display?
One certain thing is to make it SMALLER.. A good insight from Magic Leap is you can make it a pure display, without any computing power..

Okay, it is still weird. Why can't we just have the goggles like what Snapchat had:
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/7453019/bbishop_161111_1279_0012.jpg)
Make it more stylish..
And thinner <- this is probably very difficult given these many sensors around your eyes.
But wait a second, do we really have to put that many sensors around your eyes?
Make it smaller via functionality offloading to separated devices...
As I suggested earlier, make the display piece as simple as possible. Your glasses-like-or-goggles-like device should just be the display!
The necessary cameras that track hand movement and surrounding environment can be somewhere else. For example, around the neck like:

The difficulty is to sync with the cameras properly with your head movement. But I think it is totally doable, as long as the stitching process for all of the cameras is must faster than head rotation changes. Another problem is streaming display - which could be worked around if you have just a cable that goes through the display, this "collar", and the power pack.
Another hard thing is to keep the cameras in place. Well you could just have a small camera on the goggles, and align that video stream with the cameras on your neck. Even if the cameras on the neck are moved, it would be easy to adjust the scene coordinates based on the alignment video stream...
Back in 2015, I wrote about functionalities offloading, which suggests to separate functionalities on different devices, for energy saving. Now I think it makes more sense to do so for simplicity and a neat system design. You pile up functionalities for one device because it supposes to run multiple things, not because simply you can do it.
Surely it can complicate the process you wear multiple devices and route the cables to connect them, comparing with just a single headset. But considering the weight and the look, I personally appreciate much more if you just separate this bulky headset into multiple pieces.
Another unrelated, random idea coming out of nowhere
Another fun not may not be very useful idea: if you got to put eye-like cameras on the headset, can we have a "shutter" so they mimic your blinking? This would make you look even funnier, but more liveness, like a living robot. Lol.
Hmm interestingly I found NReal did very similar things I was suggesting