Imagine if you could take a picture of anything, add a little note, have it filed away. Not necessarily an awesome Instagram picture, but just a picture of some mail you got, a tool you are putting away, any thing you want to record and save. Heck, why not a picture of your computer screen? Pair that with quickly available audio transcriptions and you can also dictate anything, thoughts, small notes, information associated with the images.
That all could be great... if the library of things you made was useful. It's pretty clear how it could be useful now; do some OCR and other detection on images, use a vector store for both that and transcripts, and hook it all up to an LLM assistant. It's a bit complicated, and a bit expensive to run, but at least for a prototype you could make something pretty incredible.
Meta might make something like that... but they aren't doing that yet, and they might never do that. If the platform was open people could explore these things right now. And it doesn't even need to be radically open, you don't need to be able to hack the firmware; but it has to be more open than the preview Ray-Ban glasses, and I'm assuming more open than this revision.