I've noticed most of Google's CAPTCHAs these days are based around cars, roads and street signs. Where their implementation once allowed for assisting in the sc

I've noticed most of Google's CAPTCHAs these days are based around cars, roads and street signs. Where their implementation once allowed for assisting in the scanning of books, now it could potentially be used to funnel data into Waymo (or it could just be coincidence). I think issues like this image recognition problem show just how far away we are from realistic self driving vehicles. There have been articles on HN recently about NYCs crumbling rail infrastructure. Chicago's is no better. Only Seattle seems to be expanding theirs in the US. America needs an upgraded infrastructure because AI (in this case classification problems) have a lot of hard limits.