I'm not sure I understand the point about a dedicate "selfie" camera, however I think we're conflating "percentage of users" with "varieties of use cases." I think there could be quite a cornucopia of potential use cases, but I think per capita most people will not actually be making use of these. As other commenters have pointed out, I'd be a lot more tolerant if the data were not constantly piped to Meta.

The point about a dedicated selfie camera was that in 2008, few would have considered taking selfies to be a major use case that would drive >90% of teens and adults to have a camera which has no other reasonable purpose. In the age of FaceTime calls, it would seem absurd to question why it's needed, but nothing like that was mainstream in 2008, which would lead to the same argument of "there are very few legitimate reasons to want such a camera (and it will enable creepshots)".

My wider point is that there are already many obvious use cases, and as adoption of cameras which are always on or plausibly always on rises, there will be a lot more, including augmented reality, translation, context hinting, AI agent awareness for assistants and personal security, and at least dozens of others, some of which I am sure no one has started building for, yet.

Meta is probably not the winner in this space (or, I hope not, at least, so we agree there!). However, the idea that people have a right to remember and process what they see and hear in full fidelity is pretty basic, in my opinion.