I specifically addressed this misunderstanding in this talk: https://www.youtube.com/watch?v=Twynh6xIKUcat 38:48 while explaining this work: https://people.xiph.org/~greg/confidential_values.txt
You can think of it this way: When you sign a message you prove knoweldge of a private key (discrete log of a particular public key). Everyone can verify the signature, and yet they do not learn anything about the private key they didn't know before seeing the signature.
There is no conflict between verifyability and privacy.
It seems plausible that fully homomorphic encryption will eventually enable a practical and fully anonymous cryptocurrency, but nobody has figured out how yet. Also, even without FHE, maybe someone will figure out how to make a Bitcoin-style public-ledger system that somehow uses Chaumian blinded keys instead of ditching anonymity entirely.