As I've posted, there have been quite a few GBC competitions that were held entirely online and had none of the problems you've mentioned. Power requirements don't vary, coolness is easily gauged, reliability is assumed. It's much easier to build a functional module compared to making a video where a nonfunctional module appears functional, doing it right is easier than faking it, so it's reasonable to assume that entries aren't dropping every other ball.
Also, we're not looking to find out which one is scientifically proven to be the best module. Often, competitions have specific requirements (make a themed module, make a wheel motion, seed part challenge) and *that* is the strongest metric by which entries are judged. There's no reason to score color choices or piece counts to the hundredth of a percent
I would understand the confusion if this hadn't been done before, but it has, plenty of times. Online GBC competitions aren't new