All very good points. Don't write code, solve the problem. For that, first understand the problem. Take time to reduce complexity, else you won't be able to evolve. Gather knowledge along th he way.

This all takes a bird-eye view and a long perspective, very unlike quarter-results-driven development.