@xwu that is a straw man. Nobody is talking about shipping closed protocols with no conforming types.
Closed protocols are different in that it is less common for a library to want to ship one. But they are still a valid design technique and there is value in having a consistent design policy of reserving flexibility for library authors by default. This change may not be possible anymore but it is certainly a valid position to believe it would be better than what we currently have.