You're right, some people are misguided and think age verification can be done in a way that isn't the death of privacy and anonymity online.
Don't we all hate social media? From that standpoint, anything that makes it hard to use or come with direct negative consequences is good.
Which is true. Irma/Yivi [1] in NL/BE proves that to be the case. But it always works with a trusted third party. The client, after confirming auth with the human, generates a unique key for the platform. The platform asks the trusted third party to verify this key, and then a scope is defined. For example: 'is this person 18+?' the response to the platform is then 'Yes' or 'No'.
[1] https://yivi.app