But humans learn by themselves, from observing the world around them and drawing conclusions. Which is a major advantage and a core problem in strong AI.
The only reason it works is that a lot of humans put a lot of effort into more or less figuring out exactly how to ride a bike. It doesn't generalize at all, teaching the same robot to ride a skateboard would mean doing it all over again.
Yes, and not just hypothetically; more generally, transfer learning is definitely a thing, and has been shown to reduce training requirements for unrelated tasks.
The only reason it works is that a lot of humans put a lot of effort into more or less figuring out exactly how to ride a bike. It doesn't generalize at all, teaching the same robot to ride a skateboard would mean doing it all over again.