Agree with that! Data science is a craft and one needs to make many mistakes on the road to mastery.

By working on projects - at work, in school or just pet projects - you solve Nos 3 and 4 (everyone is better at selling own experience than knowledge).

Nos 1, 2 and 5 are also true. I'd argue not to apply for a position with a company where you are not compelled by 1 and 5.