Nothing has changed. The premise of that blog post is that we'd like to do better than ephemeral intersection requests using truncated hashes of contact identifiers, but that we don't believe it's currently possible to do better.
I don't know what you mean by "in the clear," but Signal has always done contact intersection with an ephemeral query of truncated hashes of phone numbers.
I was under the impression that this had been reverted because it is "hard". Has that decision changed?
https://whispersystems.org/blog/contact-discovery/