Explaind by Jason Smith from CeramicSpeed

I'm curious to see what mechanism they use to change gears.

It fails the KISS test.
Looks too complicated. And aluminum ain't that strong.
It's nice to look at though.

What is the KISS test? Don't know what that is a reference to?

I assume it's "Keep It Simple, Stupid"

I found the video interesting, mostly because the guy talking acknowledged that this clever tech required a redesign of half of the bike--work that hasn't been done yet.

I seriously doubt we'll ever see this on a production bike, because the engineering problems of making this work in the real world look insurmountable to me.

I thought this lends itself to a spring loaded spiral shaft automatic transmission, where the more torque applied, the more it compresses the spring towards the front crank, lowering the gear you're in. This could be adjustable, allowing a rider to select how hard they want to pedal and automatically adjusting the gears to keep that force stable at any speed.
A second gear in the rear, rotating in the opposite direction and sandwiching the drive gear, would go a long way towards stopping slippage and gear wear. They certainly need to ditch the aluminum gears, though.
Just what sprang to mind when I saw it.

Yeah, neat idea, but I'm not seeing the applications outside of (maybe) high end competitive racing where a few percent efficiency (that isn't offset by losses elsewhere, like weight) is worth it. In that case the aluminum would be okay because you'd just replace it after every event. For the likes of e-bikes and such, on the other hand, if you're going to use a shaft drive, why not just put the transmission entirely on the drive sprocket (maybe with a CVT like Netwboy suggets). This thing, as a whole, strikes me as too much out-of-the-box engineering without actually leaving the box. Of course, all that said, I'm always in favor of trying new engineering ideas so I approve.

At 4:20 of the vid i linked he shows what i think you are asking about.
Also @newtboy

The basic action is, but not the mechanism.
My idea....think spiral channels inside the tube with the cog shown attached to a piston that rides in the spiral channels. As you turn it and force is transferred, it forces the piston forward because the spiral turns rotational force into linear force. With a spring, you apply an opposing linear force so the piston only moves when those forces are unbalanced. This spring could be tunable so you select where the balance point of those forces is, thus selecting the maximum force you could apply before it changes gears for you. When there's more force applied, it "lowers" the gear, when less it automatically goes up a gear. No electronics or battery required.

I think the primary problem would be 100% of the force is transmitted to one tooth at a time, not many like a chain does. Wear will be a problem.