I would happily own a car from the 20s. Parts aren’t available, but for not much money I can buy a mill and a lathe and fabricobble together whatever I need. They’re not complicated to work on.

Well the reliability of a boot has really nothing to do with the fluid it contains. Either it has a hole in it or it doesn’t. And there’s actually no reason this couldn’t just use grease, in fact it might have to. But unlike a CV joint, this only needs to articulate in 1 axis. No rotation, no translation. Clearly this is not the complete assembly we’re seeing here however. There’s a wide range of much more fundamental mechanical problems that must be solved that this simple animation doesn’t show. There’s a number of other ways to seal an assembly like this as well. I’m very interested to get my hands on one to tear down. Maybe I’ll go patent diving and see what I find.

Just move them inboard. Might as well. You’d want to offset the additional unsprung weight, plus with the axle shaft being unreduced, you can now use much smaller and lighter brakes spinning at a higher rpm for the same effective brake torque.

With a boot.

Sure they do. That’s cause they’re not transmitting much torque at all. They don’t have to. You don’t have a lot of torque anywhere except at the ring gear - and not on the gears driving it. You’re thinking based on traditional drivetrain heuristics, where you might easily be putting upwards of 8000lbft through the axles. But as this eliminates the need for most or all pre-reduction, even direct driving off an enormous electric motor, it’s a way, way less torque than you’d see even on the cv joints of a geo metro.

It’s in ATF mode