I had a coworker that worked on the fraud team at Second Life. He said the engineers ran the place and would work on stuff that was technically interesting to them, rather than working on stuff that users wanted or needed. That probably didn't help.

What happened: nobody needed it.

Too slow (latency is inevitable), too limiting, too hard to do or show anything non-trivial.

You could do interesting things if you put it a lot of time. But few have the time to spend on unclear benefits.