Agree about the art-gallery use case, and as a guy that has been involved in a lot of product companies I think it's important that the initial demo videos show solutions to real problems.
If your initial launch marketing positions the unit as some esoteric solution to a non-problem (especially if it's for a "rich persons" non-problem), people seem more apt to forget about the product overall. Show some demo videos of it being used for REAL applications, and there is a greater chance people will remember it and/or aspire to wanting one (which helps keep you stuck in their subconscious). Nobody is going to go to work and say "did you hear about that robot you can use to browse art galleries remotely?", but they will say "did you hear about that robot that you can use to go grocery shopping?" (or something similar, I don't have a better example because I can't really think of a good use for a telepresence robot, but that's just me).
Definitely a killer product though. There are lots of very serious potential uses for something like this and it looks to be executed very well.