This article is a great explainer of the basics underlying anonymous credentials. I look forward to the promised follow-up explaining real-world examples.

The key issue however is trust. The underlying protocols may support zero-knowledge proofs. But as a user I'm unlikely to be able to inspect those underlying protocols. I need to be able to see exactly what information I'm allowing the Issuer to see. Otherwise a "correct" anonymous scheme is indistinguishable from a "bad" scheme whereby the Issue sees both my full ID and details of the Resource I wish to access. Assuming a small set of centralized Issuers, they are in a position of great power if they can see exactly who is trying to access exactly what at all times. That's the question of trust - trust in the Issuer and in the implementation, not the underlying math.

In Switzerland a digital identity like this will launch this summer and the underlying infrastructure and app is open source. And the issuer of the ID and the registry that holds and verifies credentials are separated. The protocol also isn't novel and is already used in other countries (Germany(?)).

For more information check the out technology behind it: https://www.eid.admin.ch/en/technology

This is exactly it. It is a huge issue if the authentication can trivially become non-privacy preserving in a way that is impenetrable to users.