i thought by now there would be small enough machine learning models that could be run on a dash cam that would blur out people’s faces. like OpenCV stuff or whatever Panoramax are using for their street-view imagery? but i dont know, maybe it would take more powerful CPUs to do this kind of thing?

That is one of the most stupid law I’ve ever heard.

I wonder what this means for people who already own cars like Teslas that have multiple cameras constantly recording.