I have doubts about how much load that can take. I'd think even belts can handle more. There is a reason belts aren't used everywhere. Most practical things involve loads higher than belts can take. So basically it doesn't perform the function of gears which is to transfer power in a loaded system.
Loss is a overly pedantic thing to worry about. There are going to be more losses simply in how your input energy works than would be accountable to gears. You are keeping 99% of your pennies instead of 95% when there are dollars to consider elsewhere.
I agree. Not sure how much experience you have with 3d printed gearboxes, but they tend to be very noisy. Even at light loads. In contrast this is extremely quiet.
Like you I doubt it's useful for much work, but I found it interesting. Who knows, maybe useful for high rpm and low torque applications where noise is an issue.
True. I would think though that direct electric drive would be better for that though. You don't really need gearing if you can attach a motor that achieves the right target anyway, and those can often be quite small. Like if you were trying to drive a fan (the high speed low torque application that comes to mind), it's almost certainly easier to hook it up the the motor.
It kind of reminds me of this: https://www.youtube.com/watch?v=uD7Lzv5fWhs
(post is archived)