Why should you waste your energy practicing solved problems, when there are unsolved problems to solve?

Was having this internal debate while trying to write visual SLAM from scratch, which has plenty of state of the art implementations.

I think the answer lies in the fact that solved problems are knowably solved, i.e. one knows that it is possible to solve them. And if you get stuck, there’s a non-zero amount of people that can help you solve it.

Whereas for an unsolved problem, nobody’s ever done it before. And having the conviction to even try it is ridiculous.