naertcxx point is that you shouldn't use shared_ptr for your next ptr in your list (or other node based container) node. It is slow and error prone. Instead the list container itself should own the nodes and delete them on container destruction. I think it is a good point. unique_ptr is better, but still not always ideal.

For algorithms like the ones in CLRS, which are explicitly written and proven with pointers in mind, definitely.

As I wrote, the stdlib++ agrees. Read that code and reevaluate your view on whether it is ridiculous.

For other graphs, it depends on the specific application. I am not saying that shared_ptr is always wrong, but often it is.